Food and Bioprocess Technology (2024) 17:828–849

https://doi.org/10.1007/s11947-023-03163-5

REVIEW

Intelligent Cutting in Fish Processing: Efficient, High‑quality, and Safe

Production of Fish Products
Jiaying Fu1 · Yingchao He1 · Fang Cheng1,2

Received: 20 May 2023 / Accepted: 30 June 2023 / Published online: 8 July 2023
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023

Abstract
Fish processing is an indispensable part of fish food production. It mainly involves de-heading, gutting, filleting, skinning, trimming,
and slicing, with the cutting operations holding a critical role. Unfortunately, inefficiency, low quality, and poor safety are the primary
problems facing the fish processing industry today, dramatically hindering the automation and intelligence of fish processing. Con-
sequently, it is vital to develop intelligent cutting in current fish processing in an efficient, high-quality, and safe manner. This review
summarizes the main cutting techniques for fish processing. The critical techniques to achieve intelligent cutting in fish processing
from imaging, image processing, and modeling dimensions are outlined, with their applications in practical fish processing. Fish
characteristics, cutting mechanisms, and cutting process control are emphasized. In addition, Industry 4.0 technologies, especially the
Internet of Things (IoT), big data analytics, and digital twins (DT), are emphasized. Finally, challenges and future work are highlighted,
which will serve as references for subsequent researchers and enterprises engaged in this field to promote the automation and intel-
ligence of fish processing production, ultimately realizing the high-efficiency, high-quality, and safe production of fish food products.

Keywords Cutting techniques · Image techniques · Artificial intelligence · Modeling analysis · Industry 4.0 technologies

Abbreviations FAO Food and agriculture organization of the

2D Two-dimensional united nations
3D Three-dimensional FE Finite element
ADAMS Automated Dynamic Analysis of Mechanical FRSTCA​ Fuzzy-rough set model based on the thermal-
Systems charge algorithm
AI Artificial intelligence GML Gaussian maximum likelihood
ANN Artificial neural networks GVA Gray value analysis
BSI Blade sharpness index HSI Hyperspectral imaging
CNN Convolutional neural networks IoT Internet of Things
CT Computer tomography LDA Linear discriminant analysis
DL Deep learning ML Machine learning
DT Digital twins MPa Megapascal
MRI Magnetic resonance imaging
* Fang Cheng MVS Machine vision systems
fcheng@zju.edu.cn N/A Not applicable
Jiaying Fu PCA Principal component analysis
jiayingf@zju.edu.cn RCNN Region based convolutional neural network
Yingchao He SFS Sequential forward selection
yche@zju.edu.cn SVDD Support vector data description
1
SVM Support vector machine
College of Biosystems Engineering and Food Science, TIP Traditional image processing
Zhejiang University, 866 Yuhangtang Rd., Hangzhou,
Zhejiang 310058, China UV Ultraviolet
2 VGG Visual geometry group
Key Laboratory of Intelligent Equipment and Robotics
for Agriculture of Zhejiang Province, Zhejiang University, VIS/NIR Visible and near-infrared
Hangzhou, Zhejiang 310058, China WJC Water jet cutting

13
Vol:.(1234567890)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Food and Bioprocess Technology (2024) 17:828–849
Introduction
The fish represent a highly diverse group of marine and
freshwater species, including tuna, mackerel, cod, sardine,
porgy, sole, salmon, trout, whitefish, bass, catfish, shad,
tilapia, and others, containing many nutrients the human
body needs (Arthington et al., 2016). The latest data from
the Food and Agriculture Organization of the United Nations
(FAO) showed that the gross world production from fishing
and aquaculture in 2020 stood at 178 million tons, with a
total value of approximately USD 406 billion (FAO, 2022).
Owing to its high nutritional and economic value, fish has
become one of the most important food resources. However,
the irregular, deformable, and slippery nature of fish has made
the fish processing industry labor-intensive. Fish processing
involves many steps, including receiving, bleeding, gutting,
de-heading, filleting, trimming, and slicing, with limited
cleaning and grading steps (Vidaček & Bugge, 2016). Addi-
tional steps may involve descaling and skinning. Usually, a
preliminary classification based on the species and sizes after
harvest is necessary (Dowlati et al., 2012; Sung et al., 2020).
The fish are then bled to minimize the effect of blood on the
shelf life of the flesh (Einarsdottir et al., 2022). Fish such as
salmon and trout are chilled after bleeding, which reduces
the activity of bacteria and enzymes (Chan et al., 2020). In
addition, the fish still needs to undergo gutting. Subsequent
processing can then be carried out depending on the final
product. Sometimes the fish will be whole-packaged. While in
others, they will be de-headed and even filleted, trimmed, and
further sliced to meet the diverse demands. During the whole
processing of fish, cutting plays a critical role and directly
affects efficiency and quality.
Compared with the cutting process of solid food, butcher-
ing fish is more complicated due to their contours and varia-
tions, during which the fish are gutted and further processed
into fish chunks, fillets, or strips. Sometimes up to 70% of
fish is lost when removing inedible or low-value parts in
producing fish fillets, such as viscera, head, bones, fins, and
defective parts, partially due to inefficient cutting operations
(Liu et al., 2022). In traditional fish processing, experienced
workers handle the fish with fish knives, tweezers, scrap-
ers, spatulas, and saw blades; their skills and experience
determine the efficiency and quality of the fish processing.
However, poor ergonomics and working conditions in fish
processing plants lead to health problems such as muscu-
loskeletal, skin infections, sepsis, and allergic respiratory
diseases (Jeebhay et al., 2008). As a result, it is difficult
for fish processing plants to retain experienced workers for
extended periods (Romanov et al., 2022). In recent years,
some fish processing plants have introduced automated
equipment to fish processing, significantly improving pro-
duction efficiency (Buljo & Gjerstad, 2013). However, in
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
829
some processes, especially trimming the fillets, it is much
harder to reach the level of hand cutting. Many small and
medium-sized fish processing plants still use manual trim-
ming of fillets, which dramatically increases the economic
cost. These existing problems make the research of intel-
ligent cutting in fish processing an urgent mission.
Intelligent cutting takes traditional cutting as the basis,
aiming at the automation and intelligence of fish processing
with advanced techniques. Cutting techniques are the first
concern of the research. In fish processing, methods for cut-
ting fish concentrate on metal blade cutting (Ageev et al.,
2021; Dowgiallo, 2008; Kapłonek et al., 2020). Advanced
cutting techniques, such as water jet cutting (WJC) and
ultrasonic vibration-assisted cutting (UVAC), are capable
of higher precision and quality (Gupta, 2020). UVAC can
handle sticky and elastic foods, making it possible to cut fish
(Schneider et al., 2011). It is followed by applying advanced
techniques in fish processing, including image techniques,
modeling analysis techniques, and artificial intelligence (AI)
techniques. These techniques are applied to optimize cutting
operations and guide fish processing, resulting in improved
quality and efficiency. Specific research includes fish feature
recognition, fillet defect detection, cutting parameters deter-
mination, and cutting process control. During the cutting
process of fish flesh, the flesh will undergo changes of state
from elastic deformation to plastic deformation and then
to fracture (Spagnoli et al., 2019). The shorter the elastic
and plastic deformation phase, the higher the quality of the
cut. Skin, flesh, and bones have different physical proper-
ties and require different cutting parameters. Hence, deter-
mining the optimal cutting parameters for better cut perfor-
mance is essential. In addition, the penetration of Industry
4.0 technologies in the food processing industry has led to
a focus on smart sensors, robotics, the IoT, big data analyt-
ics, and DT, which are critically important for improving
the efficiency, scalability, flexibility, and sustainability of
intelligent cutting in fish processing (Hassoun et al., 2022a;
Xu et al., 2023).
Food quality, safety assurance, high productivity, low haz-
ards, and Industry 4.0 are driving the development of intel-
ligent cutting in fish processing (Hassoun et al., 2022b, 2023;
Xu et al., 2022). Liu et al. (2022) outline the applications of
cutting techniques, machine vision, and AI in fish processing.
Xu et al. (2023) review robotization and intelligent digital
systems in the meat-cutting industry from robotic cutting,
perception, and digital development perspectives. However,
the lack of a systematic outline of intelligent cutting in fish
processing has limited the development of the fish process-
ing industry to some extent. This paper provides an overview
of intelligent cutting in different segments of fish process-
ing from the perspective of efficient, high-quality, and safe,
combining existing technological advancements (cutting
13
830
techniques, image techniques, and modeling analysis tech-
niques) and Industry 4.0 technologies (smart sensors, robot-
ics, AI, DT, big data analytics, and the IoT). Figure 1 presents
the schematic diagram of intelligent cutting in fish process-
ing. Finally, the challenges and future work are highlighted,
promising to provide new perspectives on the development
of automated and intelligent fish processing.
Cutting Techniques in Fish Processing
In fish processing, the first step is determining the cutting
techniques required for the fish or fillets in different process-
ing steps. The selection of fitted cutting tools and methods
following the features of the fish or fillets allows for less loss
of flesh and more efficient processing. Traditional cutting
is primarily based on metal cutting tools, including linear
cutting, involute cutting, and saw cutting. In the current
processing of fish, they are still the primary cutting forms.
Advanced cutting techniques, such as WJC and UVAC, are
becoming increasingly popular due to their ability to cut
more accurately and efficiently, which may bring new devel-
opments to the fish processing industry.
Metal Blade Cutting
Linear cutting and involute cutting are the most common
cutting methods. The former cuts the object through linear
motion, while the latter cuts the object through circumfer-
ential motion by attaching one end of the cutter to the motor
shaft. Planar blades are the primary cutting tool in linear
cutting (Kapłonek et al., 2020). With a unique cutter design,
linear cutting can also be used for fish bleeding, as presented
in Fig. 2a (Bonder et al., 2011). Involute cutting, by contrast,
uses a cutting tool with a specific curved or straight shape.
In fish processing, linear cutting and involute cutting are
widely used in de-heading and slicing. Figure 2b presents
Fig. 1  Schematic diagram of
intelligent cutting in the fish
processing
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Food and Bioprocess Technology (2024) 17:828–849
the automated slicing of fish fillets utilizing an involute cut-
ting strategy.
Saw cutting includes straight saw cutting and circular saw
cutting. Usually, straight saw cutting is used to produce fil-
lets and cut fish bodies (Lima et al., 2022). Figure 2c pre-
sents the robotic post-trimming system utilizing straight saw
cutting (Bar et al., 2016). Circular saw cutting has broader
applicability, involving de-heading, gutting, filleting, and
trimming. V-cutting is available by operating two circular
saws simultaneously. Azarmdel et al. (2021) developed an
automated trout processing robot that removed the fish head
by V-cutting with a double disc knife, as shown in Fig. 2d.
In addition, several studies have involved modifications to
the circular saw form. Dowgiallo and Dutkiewicz (2007)
proposed a cup-shaped circular saw capable of cutting the
head of the carp in a V-shaped path, as shown in Fig. 2e, of
which the flesh loss is close to the V-cutting with a double
circular saw (Dowgiallo, 2008). However, saw cutting will
produce a large amount of comminuted fish flesh, which
may contain fragments of bones (Lima et al., 2021).
Water Jet Cutting
WJC is increasingly being developed as a new cold-cutting
technology (Hlavacova & Geryk, 2017). Water is the pri-
mary medium used in WJC, which is green and recyclable.
Compared with metal blade cutting, WJC avoids microbial
cross-contamination and dramatically reduces the cost of
replacing major wearing parts (Cui et al., 2022). It also mini-
mizes meat loss with considerable economic benefits (Wang
& Shanmugam, 2009). WJC has a wide range of applications
in food processing (Wulfkuehler et al., 2014). Kasperowicz
et al. (2019) studied the parameters required for cutting fish
by the WJC and concluded that a complete cut of rainbow
trout leaves could be achieved at a pressure of 3.5 megapas-
cals (MPa). As a leading global supplier of advanced equip-
ment and integrated systems for the aquaculture industry,
Food and Bioprocess Technology (2024) 17:828–849
Fig. 2  Cutting techniques in fish processing. a Bleed-cutting robots
utilizing linear cutting (Bonder et al., 2011); b automated slicing of
fish fillets utilizing an involute cutting strategy, courtesy of Marel; c
robotic post-trimming system utilizing straight saw cutting (Bar et al.,
Marel designed several WJC devices for trimming trout and
salmon fillets. Figure 2f presents automated fish trimming
using a water jet knife. Additionally, the cutting power of
WJC can be effectively improved by adding non-harmful
abrasives, such as sugar, salt, starch, and ice particles, in
line with the norms of green processing as well (McGeough,
2016; Xu et al., 2022).
Ultrasonic Vibration‑Assisted Cutting
UVAC is an advanced cutting technique that allows the
knife to vibrate with high speed at a frequency of 20-40 kHz
along the cutting direction. Macroscopic feeds and micro-
scopic vibrations are primary manifestations of UVAC (Xu
et al., 2022). As a result, periodic contact and separation
fluctuations are generated between the blade and the cut-
ting object. Intermittent cutting makes UVAC capable of
reducing macro cutting forces and heat during the cutting
process with high precision and smooth cutting surfaces.
Yildiz et al. (2016) and Yildiz et al. (2019) experimentally
confirmed the cutting performance of UVAC. They further
emphasized that UVAC has good cutting performance when
processing viscoelastic materials. However, there are few
applications related to UVAC in fish processing. In addition,
UVAC has a self-cleaning function compared to traditional
cutting, which can significantly improve food safety (Xu
et al., 2023). Therefore, it is crucial to develop UVAC and
apply it in fish processing.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
831
2016); d head-cutting subsystem by V-cutting with a double disc
knife (Azarmdel et al., 2021); e cup-shaped circular saw (Dowgiallo
& Dutkiewicz, 2007); f automated fish trimming, using a water jet
knife, courtesy of Marel
To sum up, fish processing is a complex process involving
multiple cutting techniques. The improvement and expan-
sion of cutting techniques impact quality and efficiency.
The superior cutting performance demonstrated by WJC and
UVAC is an emerging opportunity for the fish processing
industry. In addition, promoting the convergence of sensing
technology and cutting tools and the convergence of robot-
ics and cutting technology is imminent. Mason et al. (2022)
overviewed mountable sensing techniques on knives. Some
studies use robots or manipulators to cut fish and fillets, with
the possibility of flexible cutting performed by robots in the
future (Bar et al., 2016; Bogue, 2008; Bonder et al., 2011).
Advanced Techniques for Intelligent Cutting
in Fish Processing
The development of imaging techniques provides new insights
into fish feature recognition and fillet defect detection but
equally complicates extracting useful information from
images of fish and fillets. Traditional image processing (TIP)
is increasingly challenging to meet the high accuracy and
efficiency required in modern fish processing. AI techniques,
especially machine learning (ML) and deep learning (DL),
can speed up image processing efficiency by forecasting and
analyzing the data obtained from imaging techniques with
the help of models, providing higher-quality cutting solutions
for fish processing. Fish processing involves specific cutting
13
832
operations that remain difficult to select the optimal cutting
parameters and determine the optimal cutting control scheme.
Hence, it is necessary to investigate the cutting process of fish
or fillets using modeling analysis techniques to understand the
intrinsic rules and optimize the cutting operations and control
process. In addition, Industry 4.0 technologies, such as smart
sensors, robotics, DT, big data analytics, and the IoT, guide
the way for current fish processing.
Imaging Techniques
Visible and Near‑Infrared Imaging
In recent years, with the development of visible and near-infra-
red (VIS/NIR) imaging techniques from two-dimensional (2D)
to three-dimensional (3D) mode, it has been widely applied in
fish processing, such as classification (Li et al., 2022), count-
ing (Li et al., 2021), quality assessment (Dowlati et al., 2012),
and mass estimation (Balaban et al., 2010). 2D VIS/NIR imag-
ing provides results based on contrast characteristics in gray-
scale or color images, which is used in identifying fish parts
and detecting fillet defects. Erikson et al. (2010) demonstrated
that 2D visible imaging could detect residual blood in the body
cavity of gutted salmon. However, 2D VIS/NIR imaging fails
to get spatial coordinate information, such as flatness and vol-
ume; it is also susceptible to light effects. By contrast, the 3D
mode is more advantageous. It can detect fast-moving objects
online, obtain their shape and contrast, and is insensitive to
small light changes. Currently, the main categories of 3D VIS/
NIR imaging techniques are binocular vision, time-of-flight,
laser triangulation, and structured light. 3D VIS/NIR imaging
techniques provide richer detail of fish and fillets, facilitat-
ing the intelligent cutting of fish. Bar et al. (2016) utilized
3D machine vision to visualize 3D-scanned salmon fillets and
proposed an algorithm for determining the caudal connective
tissue excision location, as shown in Fig. 3a. Liu et al. (2021)
used a Gocator 2150 laser displacement sensor above the raw
material for scanning to obtain the actual 3D image of the fish
for fixed-length or fixed-weight slicing. Figure 3b shows the
squid body scanning process and point cloud image.
Hyperspectral Imaging
The hyperspectral imaging (HSI) technique measures data
from hundreds of narrow spectral bands, typically smaller
than 10 nm (Wang et al., 2022; Wu & Sun, 2013). The region
covered by these spectral bands includes the visible and infra-
red regions of the electromagnetic spectra (Saberioon et al.,
2017). By observing the spectral bands absorbed by the fish
tissue, a feature spectrogram can be obtained to understand their
structure and composition (Anderssen et al., 2020). Currently,
the HSI techniques have applications in the detection of fish
products, such as fish bones (Song et al., 2020), blood detects
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Food and Bioprocess Technology (2024) 17:828–849
(Skjelvareid et al., 2017), and nematodes (Siche et al., 2016),
as shown in Fig. 3c. However, the images obtained from HSI
contain a large amount of redundant data, posing a consider-
able challenge to data analysis and mining (Cheng et al., 2017).
Therefore, it is necessary to improve the detection speed by
selecting a representative critical spectral band associated with
the detection target or using DL approaches such as convolu-
tional neural networks (CNN) (Saha & Manickavasagan, 2021).
X‑ray Imaging
X-ray imaging is a non-destructive detection technique. Pla-
nar and computer tomography (CT) are the two main types of
X-ray imaging systems (Balaban et al., 2016). With the ability
to identify internal structures, X-ray imaging techniques have
been developed for analyzing ingredients in objects that are not
detectable to the naked eye. Due to the strong penetrability and
work efficiency, the X-ray imaging technique has unique advan-
tages in detecting fish bones. Several researchers have used
X-ray imaging techniques to detect fish bone in fillets (Mery
et al., 2011; Urazoe et al., 2021; Zhong et al., 2019). Figure 3d
presents X-ray imaging with line sensors and an X-ray image
of fish. Besides fish bone detection, X-ray imaging can accu-
rately determine the apparent density of various food materials
(Kelkar et al., 2015). Therefore, it is possible to use X-ray image
data to estimate the density and, thus, the thickness of the whole
fillet. Veliyulin et al. (2011) proved that planar X-ray imag-
ing at 40 kV could estimate entire herring weights accurately.
FleXicut Salmon of Marel not only positions bones by X-ray
technique and then automatically cuts the area in which the
bones are positioned with a guided robotic water jet cutter but
also cuts portions based on the volume estimate.
Others
Other imaging techniques, such as magnetic resonance imag-
ing (MRI) and ultraviolet (UV) fluorescence, have also been
applied to the detection process. MRI allows non-destructive
detection and imaging of food structures (Kirtil & Oztop,
2016). Nakashima (2020) developed a suit of handheld mag-
netic resonance sensors to measure the fat content within
the tuna. Wang et al. (2015) detected the fish bones by UV
fluorescence imaging in cod fillets. However, the detection
rate of UV fluorescence imaging is low, with only the fish
bone on the surface of the fillet being detected.
Image Processing and Artificial Intelligence
Techniques
Traditional Image Processing
TIP techniques operate on the acquired images with three
main parts: image pre-processing, image segmentation,
Food and Bioprocess Technology (2024) 17:828–849 833

Fig. 3  Applications of imaging techniques in fish processing. 2021); c applications of HSI in fish processing sector (Hassoun
a Real-time visualization of the 3D-scanned salmon fillet and et al., 2022b); d X-ray imaging with line sensors and an X-ray
illustration of the tail reference point algorithm (Bar et al., 2016); image of fish (Urazoe et al., 2021)
b squid body scanning process and point cloud image (Liu et al.,

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

834 Food and Bioprocess Technology (2024) 17:828–849

and feature extraction (Soltani Firouz & Sardari, 2022).

Image pre-processing is to suppress unwanted distortions or
strengthen certain important features. Image pre-processing
methods include geometric transformation, pixel brightness
transformation, and filtering (Chaki & Dey, 2018). After
pre-processing the image, the image or image region is
segmented, including methods based on thresholds, edges,
and regions (Cao et al., 2020). The features of the sample
are then extracted for image recognition. Usually, feature
extraction describes the image by extracting features such
as color, shape, and texture. TIP methods have proven to
help identify fish features and detect fillet defects. Azarmdel
et al. (2019) developed an image-processing algorithm for
segmenting trout fins and determining the head and belly
cutting points through the fin position. Figure 4 presents
the whole procedure. All channels in “RGB,” “HSV,” and
“L*a*b” color spaces were tested on trout images at first.
Then they selected channel “B” of the “RGB” color space
to perform segmentation, bounding box correction, and
dimension extraction on trout images. Finally, the extrac-
tion of trout fins was achieved by an initial line determina-
tion algorithm for belly-back side detection and a secondary
line determination algorithm for head-tail and fin detection.
Bar et al. (2016) detected the caudal connective tissue to be
excised by thresholding the scattered images to find the larg-
est bright areas and extracting four caudal reference points
to guide tissue cutting, as shown in Fig. 3a.

Machine Learning

ML is a subfield of AI that includes algorithms that can

derive useful information from data and use that infor-
mation in self-learning to make reasonable classifications
or predictions (Saha & Manickavasagan, 2021). It can be
divided into supervised, unsupervised, semi-supervised,
and reinforcement learning (Zhao et al., 2021). Supervised Fig. 4  Procedure for determining the head and belly cutting points
learning is the most widely used method. Through contin- (Azarmdel et al., 2019)
uous learning and training of known data samples and out-
comes, the model extracts eigenvalues and mapping rela-
is a more sophisticated ML approach that emphasizes the
tionships and predicts the outcomes of new data, intending
constant interaction of the system with the external world
to build a concise class-label distribution model predicated
but without application in cutting fish or fillets. Up to now,
on predictor features. It is typically used for classification
ML has been widely used for fish identification, classifi-
and regression. Some standard supervised learning tools
cation, size estimation, weight estimation, and counting
are Artificial Neural Networks (ANN), Decision Trees,
(Zhao et al., 2021). Moreover, it plays a massive role in
Support Vector Machine (SVM), K-Nearest Neighbors,
fish fillet defect detection and positioning (Mery et al.,
Logistic Regression, Linear Discriminant Analysis (LDA),
2011; Sivertsen et al., 2011; Wang et al., 2015).
and Naive Bayes. Clustering and dimensionality reduction
are the two main tasks in unsupervised learning. Dimen-
sionality reduction can decrease the complexity of the data Deep Learning
and help with subsequent classification and regression.
Semi-supervised learning integrates supervised and unsu- DL has been introduced into the food field as an effec-
pervised learning to enable a combination of classifica- tive ML algorithm by analyzing images with promising
tion, regression, and clustering. Reinforcement learning results (Zhou et al., 2019). Its core is ANN. DL has shown

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Food and Bioprocess Technology (2024) 17:828–849
significant advantages in feature learning, classification,
regression, and image segmentation, which are more potent
than traditional ML. However, the interpretability of DL is
not as good as conventional ML methods (Zhu et al., 2021).
The main deep neural network models frequently used are
CNN, Recurrent Neural Networks, Auto-Encoder, Deep
Belief Networks, and Generative Adversarial Networks
(Alom et al., 2019). DL has become an essential tool in the
field of fish processing. In recent years, relevant models in
DL have been applied to the classification of fish (Baner-
jee et al., 2022), identification of fish body parts (Prasetyo
et al., 2022; Yu et al., 2020), detection of fish fillet defects
(Misimi et al., 2017), and quality assessment (Riorlykhaug
& Egeland, 2019). Misimi et al. (2017) used deep CNN to
classify cod fillets with and without blood spots with 100%
confidence in predicting all 1300 images of the test dataset.
Figure 5 visualizes the activation maps in various layers of
AlexNet. AlexNet captures intricate feature representations
based on the color properties of the images to maximize the
separability of blood spots from muscle tissue.
With the development of ANN, networks with better
performance will emerge and be applied to fish feature
recognition and fillet defect detection. However, apply-
ing DL requires extensive data and substantial computing
power to support the training. It is impractical for segments
with small data and poor hardware support. Fortunately, the
Fig. 5  Visualization of activation maps in different layers of AlexNet (Misimi et al., 2017)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
835
fish processing industry provides a large number of image
information of fish and fillets that meets this requirement,
promoting the convergence of image techniques and AI tech-
niques in the fish processing field.
Modeling Analysis Techniques
Fish Body and Flesh Modeling
For processing steps of the fish body, such as de-heading and
filleting, modeling the relationship between morphological
dimensions and the internal structure of the fish is essential.
Based on such models, adaptive cutting in fish processing
is possible. Some research based on these relational models
has achieved automatic adjustment of symmetric knife spac-
ing with fish skeleton size change during filleting, such as
Filleting Machine MS 2730 developed by Marel.
While trimming and slicing, it is essential to proceed
from the characteristics of fish flesh itself. Fish flesh
is an organic fiber material. Fiber orientation and local
defects can affect the fracture, deformation, and adhesion
properties, thus affecting the cutting process (Bremer &
Matthiesen, 2020). Several studies have focused on the
textural characteristics and rheological properties, provid-
ing a basis for creating fish flesh models (Chen & Opara,
13
836
2013; Murthy et al., 2017). In the study of Ageev et al.
(2021), the Maxwell-Thomson rheological model was used
to describe the muscle tissue of fish. He further proposed a
system of resistance affecting the tool during fish-cutting
to determine the relevance of the developed theoretical
model to the actual fish-cutting process. Chen and Opara
(2013) concluded the modeling approaches for fresh and
processed foods: the first-order reaction/kinetic model, the
Gibson-Ashby equation, the generalized Maxwell model,
the finite element (FE) method, and statistical modeling
approaches. These modeling approaches allow for further
analysis of the fracture mechanics of fish flesh. Xu et al.
(2022) introduced the FE modeling methods to solid food
cutting with the finding that the cutting process was well
simulated. Overall, modeling the fish and analyzing its
cutting process contribute to understanding its distortions,
critical stresses, and the resulting cutting forces, thus better
guiding the fish-cutting operations.
Parametric Relationship Modeling of the Cutting Process
Parametric relationship modeling of the cutting process aims
to model the relationship between cutting tool properties,
cutting parameters, and food properties to determine the
optimal cutting parameters. Generally, cutter sharpness is
the first factor to be considered, which can be quantified by
the blade sharpness index (BSI). However, the BSI is inad-
equate to estimate the cutting performance for cutting fish or
fillets due to the interference of frictional contact (Xu et al.,
2022). The quality of the blade should also be considered.
Well-manufactured blades are preferable for fish processing,
with significantly decreased frictional contact. Notably, the
cutting performance of WJC differs from the blades in that
it can be characterized by parameters such as water jet pres-
sure, traverse velocity, and nozzle diameter.
For metal blade cutting, cutting velocity and slice-to-push
ratio are the dominant parameters contributing to cutting
behavior. A minimum cutting speed exceeding the critical
crushing or bending resistance of cutting objects is usually
required to realize fracture. The slice-to-push ratio is the
ratio of the tangential velocity to the normal velocity, in
which increasing values reduce the cutting force. This find-
ing is consistent with the study of Bremer and Matthiesen
(2021) and Spagnoli et al. (2019). Moreover, the influence
of the food properties on the cutting performance should
also be considered. Generally, fish or fillets with lower tem-
peratures present stiffness and have a large elasticity mod-
ule, which would require more cutting force to cut. Bremer
and Matthiesen (2020) reviewed research on the properties
interrelated with cutting process conditions from six aspects:
slice-to-push ratio, blade sharpness, cutting speed, food tem-
perature, frictional contact, and food composition. Hence, it
is crucial to determine the relationship between the various
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Food and Bioprocess Technology (2024) 17:828–849
elements in the cutting process and form a corresponding
model to guide the cutting.
Control Modeling of the Cutting Process
Fish are complex in shape and slippery on the surface, with
the skin, tissues, and bones showing a complex textural
behavior, making fish processing extremely complicated (Xu
et al., 2023). In manual operations, workers depend on natu-
ral sensory abilities, including vision, force, smell, hearing,
and experience, to handle fish and fillets (Khodabandehloo,
2022). The convergence of multiple perceptual informa-
tion and experience allows workers to make dynamic and
adaptive cuts. This control mode also applies to controlling
the cutting process in fish processing. Various sensors pro-
vide different information, such as vision sensors providing
image information of fish and fillets, force sensors offering
continuous force feedback, and ultrasonic sensors giving dis-
tance information. By modeling the relationship between
multi-sensing information and cutting behavior and trans-
forming the cutting process into state-space equations, the
cutting process of fish and fillets will become more control-
lable. In addition, the self-learning capability possessed by
deep learning can continuously optimize the control model
and improve its robustness, rapidity, and accuracy.
The control modeling of the cutting process also involves
kinematic and dynamic modeling. Kinematics, particularly
inverse kinematics, is instrumental in solving the motion rela-
tionships of each joint of a robot or multi-axis machine. For
dynamics modeling, some software is required. Azarmdel et al.
(2021) conducted step simulation in Automated Dynamic Anal-
ysis of Mechanical Systems (ADAMS) software for dynamic
analyses of the trout-processing system. Furthermore, inter-
ference with the model from external disturbances and noise
should be avoided.
Industry 4.0 Technologies
The Internet of Things
The IoT refers to data transfer between interconnected com-
puter devices and machines (Hassoun et al., 2022a). It includes
the device layer, the network layer, and the application layer.
The application of the IoT allows for increased connectivity
and improved productivity and quality throughout food pro-
cessing. In fish processing, the interaction of data and infor-
mation usually occurs between workers and equipment and
between equipment and equipment, with data transmission
and processing determining the stability and reliability of the
production process. The IoT can fuse multi-sensor information
and environmental information for the optimal cutting of fish
or fillets online; it can also send alerts or maintenance notifica-
tions by obtaining information on the operational status of the
Food and Bioprocess Technology (2024) 17:828–849
related equipment. In addition, another importance of the IoT
is reflected in real-time traceability, safeguarding the safety of
fish products to some extent.
Big Data Analytics
Big data analytics is an advanced technology that extracts val-
uable information by processing massive amounts of data for
real-time decision-making. Unstructured data such as volume,
variety, value, velocity, and veracity are the essential character-
istics of Big data analytics (Ren et al., 2022). Fish processing
production provides a large amount of image, cut, equipment,
and environmental data. By processing these unstructured
data, big data analytics is capable of providing correspond-
ing optimal cutting solutions for the different segments in fish
processing, including cutting parameters and paths. However,
big data analytics in the fish processing industry is still in its
infancy, and the dramatic increase in data usually involves
higher cybersecurity risks (Duong et al., 2020). Hence, par-
ticular attention should be paid to the security and privacy of
the data.
Digital Twins
The DT is an advanced technology that models physical objects
or systems into digital form and establishes real-time mapping,
interaction, and communication between physical entities and
digital models (Verboven et al., 2020). Sensor data can be used
to build and control the digital model, with the digital model
controlling the physical object through the feedback system.
Technologies like the IoT and big data analytics have also con-
tributed to developing DT. DT offers superior monitoring, pre-
diction, and optimization capabilities and is increasingly being
applied as versatile digital technology in complex production
and manufacturing environments (Koulouris et al., 2021). Cur-
rently, DT has been developed in the Industrial field. However,
due to the fish body and flesh modeling difficulties and the
complicated cutting environment, DT is still at the beginning
stage in the fish processing industry.
Applications of Intelligent Cutting
in Fish Processing
Fish processing mainly involves de-heading, gutting, fil-
leting, trimming, and slicing. Other processing steps
may include bleeding, de-finning, and skinning. With the
advancement of cutting techniques, image techniques, mod-
eling analysis techniques, and Industry 4.0 technologies, the
fish processing industry is moving toward automation and
intelligence. Table 1 lists some research on intelligent cut-
ting in fish processing. Table 2 lists some intelligent process-
ing solutions for fish from BAADER and Marel.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
837
De‑heading
Generally, head removal is the first or second cutting step,
separating the high-value fillets from the relatively low-
value heads (Liu et al., 2022). For maximum yield, the de-
heading process always retains as much flesh as possible
on fillets and ensures that no part of the gills or skull is
included. Straight cuts, diagonal cuts, V-shaped cuts, and
curved cuts along the gills are the four primary forms of
head removal, with the latter two approaches resulting in less
loss of fish flesh. Disc knives and circular saws are the most
common cutting tools. An automated trout processing robot
developed by Azarmdel et al. (2021) removed the fish head
by V-cutting with a double disc knife, as shown in Fig. 6.
Determining the fish head cutting position is also important.
Currently, the location for cutting the head, including the
cutting point and cutting line, is mainly determined by the
gill position, collarbone position, pectoral fin position, and
critical points or lines around the head. The precise loca-
tion of the fish head cutting points or lines can dramatically
reduce the loss of fish flesh during de-heading. In addition,
the size and weight of the fish are also the reference data
to determine the cutting position of the head. By modeling
the relationship between the cutting position and the fish
size or weight, it is possible to develop new algorithms for
quickly determining the cutting position of the fish head.
Furthermore, inspired by the excellent image processing
performance of AI, Diamond et al. (2021) used U-Net to
segment different parts of the fish on 2D images, which pro-
vided new insights into the determination of the position for
cutting the fish head.
Gutting and Filleting
Gutting is usually done by cutting, followed by water
washing out, claw pulling out, or vacuum sucking out the
viscera, which aims to extend the shelf life and delay the
spoilage caused by the spread of bacteria. In processing
fish with high economic value, especially salmon and trout,
gutting is usually the first step to be completed. Deter-
mining the belly-cutting position is critical to completely
removing the viscera. Intelligent belly cutting requires ana-
lyzing the size and morphological characteristics of the
fish obtained by the machine vision systems (MVS) and
manipulating cutting tools such as knives or disc knives
to cut the belly based on the analysis results. Azarmdel
et al. (2021) developed a fully automated trout processing
system, as shown in Fig. 6. The belly-cutting subsystem
determines the starting position of the belly-cutting based
on the position of the anal fin. It achieves the belly cutting
from that position but is limited by the 2D image with
low cutting accuracy. In addition, some companies trim
13
838 Food and Bioprocess Technology (2024) 17:828–849

Table 1  Research on intelligent cutting in fish processing

Applications Objects Cutting techniques Advanced techniques Reference

Bleeding Salmon and trout Linear cutting • A line‐scan laser triangulation system for image (Bonder et al., 2011)
capture;
• TIP methods for 3D segmentation;
• LDA classifier for head–tail classification and
incision-point localization
De-heading and gutting Trout Circular saw cut- • A video camera (Basler) for image capture; (Azarmdel et al., 2021)
ting (V-cutting) • TIP for determining the head-cutting and belly-cut-
ting point according to the pectoral and anal fins;
• Modeling in ADAMS for dynamic analyses of the
trout-processing system
Trimming Salmon fillets Robotic cutting • A 3D line scan camera for image capture; (Bar et al., 2016)
• TIP for determining the tail-cutting reference points
Slicing Squid Linear cutting • A laser displacement sensor for scanning the raw (Liu et al., 2021)
material;
• A simulated annealing algorithm for the multi-
objective cutting optimization problem of irregular
fish-body shapes
Slicing Frozen fish Linear cutting • 3D laser line profile sensors for building 3D fish (Lopez et al., 2019)
models;
• Modeling for determining the relationship of weight,
volume, and density of fish to conduct fixed-length
and fixed-weight cutting

the bellies of fish that have been gutted. Filleting Machine
MS 2730, developed by Marel, trims the belly of the gutted
salmon with a rotary disc knife before filleting.
Filleting is a unit operation done in the preliminary fish
processing, in which the fish is cut lengthwise to produce
fillets (Jayraj et al., 2019). Fillets are fish flesh composed
of dorsal and abdominal muscles separated from the back-
bone and fins. Most companies that commercialize the
production of fillets use filleting machines that cut along
the spine, using a symmetrical pair of knives to cut the
ribs and vertebrae (Borderias & Sanchez-Alonso, 2011).
The predominant cutting technique used in filleting is saw
cutting, which includes straight and circular saw cutting
(Jayraj et al., 2019; Lima et al., 2022; Ørnholt-Johansson
et al., 2017). However, the presence of the fish backbone
makes the filleting process one of the processes with the
highest rates of flesh loss, although the residual flesh on
the fish backbone can be recycled and processed into other
fish products (Abdollahi et al., 2021; Rustad et al., 2011).
Related studies have found that the variation of the trans-
verse diameter of the fish backbone along the spine line is
determined by the morphology characteristics of the fish

Table 2  Intelligent processing solutions for fish from BAADER and Marel

Manufacturers Device model Applications Throughput Product advantages

BAADERa BAADER 581 Pro Filleting • Up to 25 fish/min • Dynamic cutting patterns

• 2–7 kg • Suitable for a range of different species
• Easy to adjust and maintain
BAADER 52 Neo Skinning • Up to 150 fillets/min • Shallow and deep skinning possible
Marelb I-Cut 130 PortionCutter Slicing • Up to 1000 cuts/min • High-accuracy cutting minimizes waste
• Easy-to-clean design maximizes hygiene
Flexicut Salmon Trimming • Up to 36 fillets/min • Increased high-value product ratio
• X-ray bone detection
• Precision cutting with WJC
Filleting Machine MS 2730 Filleting • Up to 25 fish/min • Automatic infeed and trimming
• 1.5–10 kg • Dynamic cutting patterns
• Sturdy and hygienic design for maxi-
mum uptime and easy cleaning
a
https://​www.​baader.​com/
b
https://​www.​marel.​com/

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Food and Bioprocess Technology (2024) 17:828–849
Fig. 6  Fully automated trout
processing system includes a
head-cutting subsystem and
a belly-cutting subsystem
(Azarmdel et al., 2021)
body. Investigating the fish backbone model makes it feasi-
ble to control the symmetric blades along a predetermined
trajectory. BAADER 581 Pro, developed by BAADER, has
achieved the dynamic cutting process along the fish back-
bone. In addition, the filleting performance of the filleting
machines is readily influenced by the blocking problems of
the blades due to some fish offal; this leads to the fact that
a certain amount of labor is required to clean and maintain
machines, which raises the costs (Liu et al., 2022).
Trimming
Defects and residues may be present in the fillets due to the
fish itself or the machine. Pin bones are the first concern in
producing the desired boneless fish products (Skjelvareid
et al., 2017). Other defects, such as blood defects, melanin
spots, gaps, and nematodes, represent a quality problem of
fish, affecting food quality and safety. Usually, pre-harvest
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
839
stress lead to rapid onset and extreme stiffness, potentially
resulting in gaping and bad bleeding (Olsen et al., 2008).
Simultaneously, the increased blood content in the muscle
has a pro-oxidant effect, which may cause rancidity and
thus affect changes in color and fishy taste during storage
and processing (Liu et al., 2013). Fillets with melanin spots
cannot be sold in whole. Additionally, consuming simply-
processed or undercooked fish risks foodborne parasitic
zoonosis due to possible parasite infection (Simat et al.,
2015). Hence, trimming fillets to remove defects and rem-
nants is essential.
Rotary disc knives are the primary cutting tools used in
filleting (Ørnholt-Johansson et al., 2017). WJC technique,
as a new food-cutting technique, has been applied to trim
fish fillets. Its excellent mobility and flexibility allow for
cutting behavior in all directions and complex curved lines
(Liu et al., 2019). The corresponding efforts were carried out
to determine a series of parameters required, such as water
13
840
jet pressure and nozzle diameter (Kasperowicz et al., 2019).
FleXicut Salmon, developed by Marel, can locate pin bones
by X-ray imaging technique and automatically cut the areas
where the bones are positioned with a guided robotic water
jet cutter. However, the WJC technique requires considerable
water and air pressure to control, resulting in high costs.
Introducing the WJC technique in small and medium-sized
fish processing plants is challenging.
Sorting the defective fillets and determining the defect
location are other priorities to consider now that the cut-
ting techniques have been settled. Table 3 summarizes the
relevant studies and works on fish fillet defect detection.
The primary image information acquisition source is the
MVS based on VIS/NIR, HIS, and X-ray imaging tech-
niques. X-ray imaging techniques have gradually become
mainstream in detecting pin bones because of their strong
penetrability and work efficiency (Song et al., 2020). For
defects visible to the naked eye, such as blood defects, mela-
nin spots, and gaps, VIS/NIR imaging techniques play an
essential role (Balaban et al., 2010; Kogiannou et al., 2022;
Misimi et al., 2017; Ooi et al., 2022). While for defects that
are not visible to the naked eye, such as nematodes, HSI
techniques are even more advantageous (Sivertsen et al.,
2012). Other imaging techniques, such as UV fluorescence
imaging, have applications in detecting pin bones and nema-
todes (Wang et al., 2015; Yang et al., 2013). In addition,
the application of AI techniques in image processing makes
the detection of fish fillet defects and residues more accu-
rate and faster. Urazoe et al. (2021) proposed a new fish-
bone detection approach using a CNN and synthetic image
generation, which can effectively detect pin bones in fillets
through synthetic virtual fish bones. Xu and Sun (2018)
detected the salmon muscle gaps using CNN features by
computer vision. With the help of DL, the task was simpli-
fied to manually labeling a set of images and training the
algorithm for identifying gap areas, significantly improving
the work efficiency. Figure 7 visualizes the final classifica-
tion result obtained from the simplified CNN model.
When the target defects or residual parts are identified and
positioned, one or more cutting solutions will be calculated
and conducted. Generally, the trimming task of fillets can be
divided into two parts: marginal trimming and optimal slit-
ting. The irregular edge of fillets, fat parts, marginal defects,
residual skins, and black films left on the edge of the fil-
lets are commonly removed by marginal trimming. In other
words, the unsightly parts of the fish fillet are cut off without
damaging the whole fillet. Optimal slitting is usually applied
for defects, such as pin bones, blood defects, melanin spots,
and nematodes. By cutting the fillets into multiple pieces,
the flawless parts of the fillets are retained, and the defective
parts are abandoned or recycled. In addition, slitting accord-
ing to the differences in the economic value of the different
parts of the fish fillet is another aspect of optimal slitting. Liu
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Food and Bioprocess Technology (2024) 17:828–849
et al. (2022) divided cod fillets into loin, center cuts, belly
flaps, and tails, indicating that the loin is usually the most
valuable part of cod fillets. Sivertsen et al. (2009) presented
a new ridge detection approach to detect the centerline of cod
fillets based on imaging spectral data, which helps determine
the optimal slitting scheme.
Slicing
Slicing is cutting the headed and gutted fish, including
chilled and frozen fish bodies, into fish steaks or cutting the
fillets into smaller slices (Ross & Edwards, 2015). Linear and
involute cutting with slicing knives is common. Bremer and
Matthiesen (2021) researched the cutting process of fish flesh
to determine a series of parameters required for slicing, such
as slice-to-push ratio, cutting speed, and cutting force. WJC
technique currently has no relevant applications in slicing due
to its small cutting area of action. Additionally, producing
fish steaks and slices with the same thickness or weight is a
critical issue in slicing. Compared with fixed-length cutting,
it is more complicated to realize the fixed-weight cutting due
to the relation between the mass and volume of the fish or
fillets. Up to now, weight and volume estimation have been
applied in fixed-length and fixed-weight cutting solutions
for fish and fillets. 3D VIS/NIR imaging and X-ray imaging
techniques are becoming increasingly practical methods for
estimating food volume, quality, and density (Kelkar et al.,
2015; Uyar & Erdogdu, 2009; Veliyulin et al., 2011). Fur-
thermore, appropriate cutting parameters should be deter-
mined to achieve fast slicing with low energy consumption
(Schuldt et al., 2018; Xu et al., 2022).
While slicing fillets, one side of the fillet is attached
closely to the conveyor belt or processing platform. It is
uncomplicated for 3D VIS/NIR imaging or X-ray imaging
techniques to obtain a 3D model of the fillet by measuring
the distance from the visible surface of the fillet to the sur-
face of the conveyor belt or the processing platform. Then,
the weight of the fillet is acquired to construct the relation-
ship between weight and volume for subsequent fixed-
weight cutting of the fillet. However, irregular objects such
as chilled and frozen fish bodies must be imaged with mul-
tiple imaging sources from different directions. Lopez et al.
(2019) set two 3D laser line profile sensors in an opposite
layout to scan the fish body, resulting in two line profiles.
The computed profiles were then integrated, generating a
precise 3D fish model using a computer vision algorithm.
In addition, the different densities of different parts of the
fish body will also hinder the mass estimation, leading to a
decline in accuracy. In recent years, ML has been used to
study the relationship between mass and volume. Liu et al.
(2021) studied the application of the improved simulated
annealing algorithm to the multi-objective optimization and
cutting problem of irregular fish in the aquatic field. An
Food and Bioprocess Technology (2024) 17:828–849 841

Table 3  Detection of fillet defects in trimming

Defects Objects Imaging techniques Image processing Results References
methods

Pin bones Salmon fillets VIS/NIR imaging Faster- Neural network (Xie et al., 2021)
RCNN + AlexNet, models can readily
Faster-RCNN + VGG detect pin bones in
fish flesh. Faster-
RCNN + VGG16
composite model
performs best
Grass carp fillets HSI FRSTCA, SVDD The method can detect (Song et al., 2020)
fish bones up to
2.5 mm in depth and
obtain a detection per-
formance of 90.5%
Salmon and trout fillets X-ray imaging SFS, SVM The detection per- (Mery et al., 2011)
formance is 100%,
98.5%, and 93.5%
for large, medium,
and small fish bones,
respectively
Tilapia fillets X-ray imaging N/A The X-ray inspection (Zhong et al., 2019)
system can quickly
detect the bones in
tilapia fillets, and
the obtained X-ray
images can be
display-delayed
Salmon and trout fillets X-ray imaging (CT) Textural analysis The length of pin bones (Sarah et al., 2019)
in trout and salmon
was measured, then a
texture analyzer and
µCT X-ray were used
to measure pulling
force, breakpoint, and
volume
Mackerel fillets X-ray imaging CNN The mean F-value of (Urazoe et al., 2021)
the CNN trained with
the 120 synthetic
images is higher
than that of the CNN
trained with ten
images of actual fish
bones. The method
can effectively detect
pin bones
Cod fillets UV fluorescence imag- LDA The detection perfor- (Wang et al., 2015)
ing mance is 90.86%
for pin bones in cod
fillets
Blood defects Cod fillets VIS/NIR imaging CNN, SVM The accuracy of accu- (Misimi et al., 2017)
rate segmentation and
localization of blood
spots is 96%
Cod fillets HSI N/A This research proposed (Skjelvareid et al., 2017)
a spectral unmixing
approach for detect-
ing blood in diffuse
reflectance hyper-
spectral images of cod
fillets

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

842 Food and Bioprocess Technology (2024) 17:828–849

Table 3  (continued)
Defects Objects Imaging techniques Image processing Results References
methods

Melanin spots Sand flathead fish fillets VIS/NIR imaging TIP This method can grade (Ooi et al., 2022)
the degree of melanin.
At the threshold value
of 90, the melanized
muscle surface area
percentage is deter-
mined objectively
Gaps Salmon fillets VIS/NIR imaging TIP The side-lighting (Balaban et al., 2011)
method was applied.
The gaps in salmon
fillets can be effec-
tively detected
Salmon fillets VIS/NIR imaging TIP A fully automated (Merkin et al., 2013)
image analysis
method is developed
for quantifying gaps
in smoked salmon
fillets
Salmon fillets VIS/NIR imaging CNN, SVM The method can be well (Xu & Sun, 2018)
applied in detecting
and classifying gaps
in salmon fillets
Sparus aurata and VIS/NIR imaging TIP The accuracy in detect- (Kogiannou et al., 2022)
Pagrus major fillets ing fillet gaps and
classifying the fillets
is more than 95%
Cod fillets HSI LDA The overall detection (Sivertsen et al., 2011)
rate for all nematodes
is 58%, with 46% and
71% for pale and dark
nematodes, respec-
tively
Cod fillets HSI GML The detection perfor- (Sivertsen et al., 2012)
mance is 70.8% and
60.3% for the dark
and pale nematodes
Cod fillets UV fluorescence imag- PCA, GVA The overall detection (Yang et al., 2013)
ing ratio using cod fillets
as samples was more
excellent than 80%

FRSTCA​Fuzzy-rough set model based on the thermal-charge algorithm, GML Gaussian maximum likelihood, GVA gray value analysis, N/A not
applicable, PCA principal component analysis, RCNN region based convolutional neural network, SFS sequential forward selection, SVDD sup-
port vector data description, VGG visual geometry group

information guidance strategy was used to optimally gener-
ate the objective solution to optimize fish weight and shape.
Others
Other cutting processes include bleeding, skinning, and fin
removal. Fish, such as salmon and trout, are slaughtered with
sharp knives by bleeding the gills of fish, which can signifi-
cantly reduce both reactive molecules and hemoglobin and
prevent the formation of blood spots in the flesh (Secci &
Parisi, 2016). Bonder et al. (2011) proposed an approach
to the salmon slaughter line that primarily used a line-scan
laser triangulation system and a blood-cutting robot equipped
with four dual-axis pneumatic actuators to treat the slaughter
of pre-anesthetized salmon, reducing manual labor require-
ments in salmon slaughter plants. Additionally, fin removal
is necessary for most fish processing, including removing
pectoral, anal, pelvic, dorsal, and caudal fins (Azarmdel
et al., 2019; Larouche et al., 2015). Filleting Machine MS
2730, developed by Marel, cuts the dorsal fin of salmon with
double disc knives through V-cutting in filleting. Hao et al.
(2022) and Yu et al. (2020) implemented the segmentation

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Food and Bioprocess Technology (2024) 17:828–849
Fig. 7  Visualization of the final
classification result obtained
from the simplified CNN model
(Xu & Sun, 2018)
of fishtail and fish bodies from TIP and DL methods, respec-
tively. While in the skin removal process, the mechanism of
most fish skinning machines is to press the fish fillet against
the blade and remove the skin from the fillet (Zieliński et al.,
2018). BAADER 52 Neo, developed by BAADER, used a
drum to press the fillet into the blades for skinning. In addi-
tion, skinning can be divided into shallow skinning and deep
skinning. Deep skinning can give the fillet less fat, a more
uniform color, a better appearance, and a higher value (Bland
et al., 2021).
Challenges and Future Work
Fish processing is an essential part of fish production, with
numerous processing equipment having been developed.
Table 4 summarizes the global manufacturers of fish pro-
cessing equipment.
The variety of fish with complex shapes, slippery sur-
faces, and textural characteristics make fish processing as
complicated as livestock and poultry processing. Simulta-
neously, the shortage of labor and skilled workers due to
health problems, such as musculoskeletal problems, poses
a significant challenge to fish processing. As a result, there
is a strong emphasis on intelligent cutting in fish process-
ing to reduce the dependence on skilled labor. In order to
catch up with the wave of Industry 4.0, potential directions
for future research on intelligent cutting in fish processing
are summarized:
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
843
1. Expansion of fish-cutting objects. Much of the research
on intelligent cutting has focused on high-value marine
fish, such as salmon, trout, cod, and mackerel, with a
need for more research on freshwater fish. Potential
freshwater fish may include catfish, blackfish, tilapia,
and others. Notably, freshwater fish tend to have more
intermuscular bone spurs, placing higher technical
demands on cutting. With the development of freshwater
aquaculture, research into freshwater fish-related cutting
will bring economic benefits.
2. Innovation of cutting techniques in fish processing. The
emergence of novel cutting techniques usually brings
dramatic changes to an industry, such as the develop-
ment of WJC has made non-contact fillet trimming via-
ble. WJC has limitations in fish processing. Due to its
small cutting surface, WJC fails to reach the efficiency
of metal blades in slicing. Therefore, further develop-
ment of WJC is the focus of our attention. WJC with
food-grade non-hazardous abrasives, such as sugar, salt,
starch, and ice particles, is a future research direction.
UVAC, as an advanced cutting technique, its inter-
mittent cutting property allows for higher quality and
precision. Meanwhile, well performance of UVAC in
cutting viscoelastic objects proves its potential in fish
processing. Additionally, robotic cutting is also a future
development direction. Furthermore, innovations in the
existing cutting tools should also be supported, such as
the research of intelligent cutting tools equipped with
sensors (Mason et al., 2022).
13
844 Food and Bioprocess Technology (2024) 17:828–849

Table 4  The global manufacturers of fish processing equipment

Manufacturers Fish species Cutting operations Fish machinery Country

Arenco Whitefish, pelagics De-heading, filleting, skinning VMK Series Sweden

BAADER Salmonids, whitefish, pelagics, De-heading, gutting, filleting, slit- BAADER Series Germany
aquaculture, tuna ting, skinning, trimming, slicing
Cretel Salmonids, whitefish, pelagics, Skinning F460A-New Skinner Generation, Belgium
aquaculture 460 TAC​
JOSMAR Salmonids, whitefish, pelagics, De-heading, filleting, slitting, JM Series Spain
aquaculture, tuna skinning
Kaj Olesen Salmonids, tuna Skinning, trimming, slicing FISKK Fish Skinner Series, Easy- Denmark
Matic Series, KSL 950 Slicing
Machine
Kroma Salmonids, pelagics, aquaculture De-heading, gutting, filleting HEADMASTER, GUTMASTER Denmark
Series, FILETMASTER
Marel Salmonids, whitefish De-heading, filleting, trimming, Salmon Deheader MS 2721 V, Iceland
portioning Filleting Machine MS 2750,
FleXicut Salmon, I-Cut 11
PortionCutter
MARELEC Salmonids, whitefish, pelagics, Portioning Portion Cutter PORTIO Series Belgium
tuna
Nikko Salmonids, whitefish, pelagics, De-heading, gutting, filleting Autoheader with hizcut, Gutter Japan
aquaculture Machine, Band Fisher
NOCK Salmonids, whitefish, aquaculture Skinning Cortex Series Germany
Norbech Salmonids, whitefish, aquaculture Gutting, filleting, skinning, trim- Gutting Machine 3000, Filleting Denmark
ming, slicing machine FR Series, Skinning
machine V Series, Pin-bone
Remover 400, Fresh slicers SL
Series
PERUZA Pelagics De-heading, filleting Automatic small fish filleting Latvia
machine–PASFF 110
Pisces Fish Machinery Salmonids, whitefish, pelagics, De-heading, gutting, filleting, HV Series, SGA-32 Head on Gut- USA
aquaculture trimming ting Machine with Roe Saving,
FR Series, Model Series
ROBOTNOR Salmonids Trimming RoboTrim Norway
Ryco Equipment Salmonids, whitefish, pelagics De-heading, filleting, trimming #231 Fish Header, #656 FIL- USA
LETER, #659 Pinboner
Salmco Technik Salmonids Slicing Salmco-Model SM Series Germany
SEAC AB Whitefish, pelagics De-heading, filleting FPM Series Sweden
Skaginn 3X Whitefish, pelagics, aquaculture Skinning Sub-Zero Skinner Iceland
STEEN Pelagics De-heading, gutting, skinning ST Series Belgium
TOYO SUISAN KIKAI Salmonids, whitefish, pelagics, De-heading, filleting, skinning, TOYO Series Japan
aquaculture, tuna portioning
VARLET Salmonids, whitefish, pelagics Skinning V Series France
Velfag Whitefish De-heading, filleting, skinning M Series Iceland

3. Convergence of image techniques and AI techniques in
fish processing. Given the variability of each fish, 2D
VIS/NIR imaging techniques are no longer sufficient to
guide accurate cutting operations. 3D VIS/NIR imaging
techniques are increasingly gaining dominance, includ-
ing binocular vision, time-of-flight, laser triangulation,
and structured light, which can provide more detailed
information about fish or fillets. Other imaging tech-
niques, such as HSI, X-ray imaging, and UV fluores-
cence imaging, provide a new platform for detecting fil-
let defects. However, the advances in imaging techniques
place higher demands on information extraction. Up to
now, AI techniques have been significantly developed
in the field of image processing. Following image pre-
processing, image segmentation, and feature extraction,
the recognition efficiency and accuracy of the extracted
features placed under a specific algorithmic model for
training far exceed those of TIP methods. With the
development of ML, especially DL, networks and mod-
els with better performance will emerge and gradually

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Food and Bioprocess Technology (2024) 17:828–849
be applied to fish feature recognition and fillet defect
detection, guiding the fish-cutting operations.
4. Exploration of the fish-cutting mechanisms. Modeling
analysis techniques are practical tools used to analyze
the fish-cutting process. Modeling and analyzing the
skeletal structure provide a theoretical basis for deter-
mining the cutting angle and adjusting the blade posi-
tion. Fish flesh material models facilitate the simula-
tion and analysis of the cutting process to determine the
required cutting parameters. In addition, more emphasis
should be on the research relating to the cutting param-
eters and their relationships with each other in fish
processing. The cutting parameters of the metal blade
include blade sharpness, frictional contact, slice-to-push
ratio, cutting velocity, and others. In contrast, WJC is
more concerned with water jet pressure, traverse veloc-
ity, and nozzle diameter. Moreover, a system to evaluate
the quality of fish and fillet cuts is essential. Apart from
determining the grade of cutting quality, judging criteria
should also be given, such as the flatness of the cutting
surface and the loss rate of fish flesh. The same applies
to judging the cutting quality of WJC. With the develop-
ment of intelligent cutting, it is urgent to have explicit
knowledge of fish-cutting mechanisms for selecting
the appropriate process parameters to guide the cutting
operations and reduce energy consumption.
5. Fusion of multi-source perception. Since cutting in fish
processing is a complex operation, a single sensor cannot
meet the complexity of the process. Multi-sensor sys-
tems are capable of compensating for the limitations of
a single sensor and providing more information to assist
in cutting. For automated large fish processing, it is par-
ticularly vital to obtain vision sensor and force sensor
information simultaneously. The convergence of vision
and force feedback allows for better flexibility and adapt-
ability in the cutting process. In addition, the accuracy
and robustness of the sensors need to be improved to
cope with the effects of unstructured environments.
6. Implementation of advanced control strategies in fish
processing equipment. Fish processing equipment is the
vehicle for performing fish-cutting operations, with con-
trol strategies determining efficiency and quality. Explor-
ing advanced analysis algorithms and control strategies
to achieve adaptive cutting is indispensable to coping
with complex fish processing environments. The com-
pensation system minimizes the impact of various dis-
turbances on the fish-cutting operations (Xu et al., 2023).
In addition, AI techniques, including ML and DL, allow
for improved decision-making through model training.
Compared to traditional control strategies, advanced
control strategies incorporate multiple new theories
and techniques that significantly improve the efficiency
and robustness of fish processing equipment and pro-
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
845
vide optimal feedback control. Therefore, promoting the
applications of advanced control strategies in the devel-
opment of fish processing equipment is essential.
7. Promoting the application of Industry 4.0 technologies
in fish processing. Industry 4.0 has brought many tech-
nologies, especially the IoT, big data analytics, and DT,
to drive the development of food processing. However,
the application of these technologies in the fish process-
ing industry is rare. The IoT and big data analytics can
provide optimal cutting solutions for fish processing,
while their real-time traceability ensures the safety of
fish products. DT can guide and optimize fish-cutting
operations by building virtual models corresponding to
physical entities. Notably, these technologies still need
to be further developed and applied to promote auto-
mated and intelligent fish processing better.
Conclusion
Fish-cutting intelligence is an inevitable trend in the fish pro-
cessing industry. This paper comprehensively reviews intelli-
gent cutting in current fish processing. The primary cutting
techniques in fish processing are introduced, including linear
cutting, involute cutting, saw cutting, and WJC. Other novel
cutting techniques, such as UVAC, may be applied to fish pro-
cessing with dramatic impact in the future. This paper also out-
lines the technological advances to achieve intelligent cutting in
fish processing from imaging, image processing, and modeling
dimensions, with their applications in practical fish processing.
Among them, applying AI techniques, especially DL, compen-
sates for the weakness of TIP in obtaining critical information
on fish images, thus improving the quality and efficiency of
cutting. Besides, corresponding intelligent cutting solutions
are given for different steps in fish processing, especially de-
heading, gutting, filleting, trimming, and slicing. In addition,
Industry 4.0 technologies, especially the IoT, big data analytics,
and DT, are emphasized. For the future development of the
fish processing industry, more attention should be focused on
the innovation of cutting techniques, the application of image
processing with AI, the exploration of fish-cutting mechanisms,
the fusion of multi-source perception, the implementation of
advanced control strategies, and the application of Industry 4.0
technologies, which will accelerate the automation and intelli-
gence of fish processing and ultimately realize high-efficiency,
high-quality, and safe production of fish products.
Author Contribution Jiaying Fu: Conceptualization; Methodology;
Investigation; Writing—original draft; and Writing—review & edit-
ing. Yingchao He: Conceptualization; Methodology; Investigation; and
Writing—review & editing. Fang Cheng: Conceptualization; Method-
ology; Resources; Funding acquisition; and Writing—review & edit-
ing; Supervision.
13
846
Funding This work was supported by the Zhejiang Provincial Natural
Science Foundation of China [Grant No. LZ23C200005].
Declarations
Competing Interests The authors declare no competing interests.
References
Abdollahi, M., Wu, H., & Undeland, I. (2021). Impact of process-
ing technology on macro- and micronutrient profile of protein-
enriched products from fish backbones. Foods. https://​doi.​org/​
10.​3390/​foods​10050​950
Ageev, O. V., Dowgiałło, A., Sterczyńska, M., Piepiórka-Stepuk, J.,
Giurgiulescu, L., Janowicz, M., & Jakubowski, M. (2021). Exper-
imental characterization and theoretical modeling of fracture and
friction resistance forces during tuna cutting. Journal of Food
Engineering. https://​doi.​org/​10.​1016/j.​jfood​eng.​2021.​110648
Alom, M. Z., Taha, T. M., Yakopcic, C., Westberg, S., Sidike, P., Nasrin,
M. S., Hasan, M., Van Essen, B. C., Awwal, A. A. S., & Asari,
V. K. (2019). A state-of-the-art survey on deep learning theory
and architectures. Electronics, 8(3), 292. https://​doi.​org/1​ 0.3​ 390/​
elect​ronic​s8030​292
Anderssen, K. E., Stormo, S. K., Skara, T., Skjelvareid, M. H., & Heia,
K. (2020). Predicting liquid loss of frozen and thawed cod from
hyperspectral imaging. LWT-Food Science and Technology.
https://​doi.​org/​10.​1016/j.​lwt.​2020.​110093
Arthington, A. H., Dulvy, N. K., Gladstone, W., & Winfield, I. J.
(2016). Fish conservation in freshwater and marine realms: Sta-
tus, threats and management. Aquatic Conservation-Marine and
Freshwater Ecosystems, 26(5), 838–857. https://d​ oi.o​ rg/1​ 0.1​ 002/​
aqc.​2712
Azarmdel, H., Mohtasebi, S. S., Jafari, A., & Rosado Muñoz, A.
(2019). Developing an orientation and cutting point determina-
tion algorithm for a trout fish processing system using machine
vision. Computers and Electronics in Agriculture, 162, 613–629.
https://​doi.​org/​10.​1016/j.​compag.​2019.​05.​005
Azarmdel, H., Mohtasebi, S. S., Jafary, A., Behfar, H., & Rosado
Muñoz, A. (2021). Design and simulation of a vision-based
automatic trout fish-processing robot. Applied Sciences-Basel.
https://​doi.​org/​10.​3390/​app11​125602
Balaban, M. O., Misimi, E., & Ayvaz, Z. (2016). Quality evaluation
of seafoods. In D.-W. Sun (Ed.), Computer Vision Technology
for Food Quality Evaluation (2nd ed., pp. 243–270). Academic
Press, San Diego. https://​doi.​org/​10.​1016/​B978-0-​12-​802232-​
0.​00010-4
Balaban, M. O., Sengor, G. F. U., Gil Soriano, M., & Guillen Ruiz, E.
(2010). Using image analysis to predict the weight of Alaskan
salmon of different species. Journal of Food Science, 75(3),
E157–E162. https://​doi.​org/​10.​1111/j.​1750-​3841.​2010.​01522.x
Balaban, M. O., Sengor, G. F. U., Soriano, M. G., & Ruiz, E. G.
(2011). Quantification of gaping, bruising, and blood spots in
Salmon fillets using image analysis. Journal of Food Science,
76(3), E291–E297. https://​doi.​org/​10.​1111/j.​1750-​3841.​2011.​
02060.x
Banerjee, A., Das, A., Behra, S., Bhattacharjee, D., Srinivasan, N. T.,
Nasipuri, M., & Das, N. (2022). Carp-DCAE: Deep convolu-
tional autoencoder for carp fish classification. Computers and
Electronics in Agriculture. https://​doi.​org/​10.​1016/j.​compag.​
2022.​106810
Bar, E., Mathiassen, J. R., Eilertsen, A., Mugaas, T., Misimi, E., Linnerud,
Å. S., Salomonsen, C., & Westavik, H. (2016). Towards robotic
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Food and Bioprocess Technology (2024) 17:828–849
post-trimming of salmon fillets. Industrial Robot-the International
Journal of Robotics Research and Application, 43(4), 421–428.
https://​doi.​org/​10.​1108/​ir-​11-​2015-​0205
Bland, J. M., Grimm, C. C., Bechtel, P. J., Deb, U., & Dey, M. M. (2021).
Proximate composition and nutritional attributes of ready-to-cook
catfish products. Foods. https://​doi.​org/​10.​3390/​foods​10112​716
Bogue, R. (2008). Cutting robots: A review of technologies and appli-
cations. Industrial Robot-the International Journal of Robotics
Research and Application, 35(5), 390–396. https://​doi.​org/​10.​
1108/​01439​91081​08935​54
Bonder, M. S., Mathiassen, J. R., Vebenstad, P. A., Misimi, E., Bar,
E. M. S., Toldnes, B., & Ostvik, S. O. (2011). An automated
salmonid slaughter line using machine vision. Industrial Robot-
the International Journal of Robotics Research and Application,
38(4), 399–405. https://​doi.​org/​10.​1108/​01439​91111​11320​85
Borderias, A. J., & Sanchez-Alonso, I. (2011). First processing steps
and the quality of wild and farmed fish. Journal of Food Science,
76(1), R1–R5. https://d​ oi.o​ rg/1​ 0.1​ 111/j.1​ 750-3​ 841.2​ 010.0​ 1900.x
Bremer, F., & Matthiesen, S. (2020). A review of research relating to
food slicing in industrial applications. Journal of Food Engineer-
ing. https://​doi.​org/​10.​1016/j.​jfood​eng.​2019.​109735
Bremer, F., & Matthiesen, S. (2021). High-speed cutting with involute
blades: Experimental research on cutting forces. Journal of Food
Engineering. https://​doi.​org/​10.​1016/j.​jfood​eng.​2020.​110380
Buljo, J. O., & Gjerstad, T. B. (2013). Robotics and automation in sea-
food processing. In D. G. Caldwell (Ed.), Robotics and Automa-
tion in the Food Industry (1st ed., pp. 354–384). Woodhead Pub-
lishing, Oxford. https://​doi.​org/​10.​1533/​97808​57095​763.2.​354
Cao, W., Liu, Q., & He, Z. (2020). Review of pavement defect detection
methods. IEEE Access, 8, 14531–14544. https://d​ oi.o​ rg/1​ 0.1​ 109/​
ACCESS.​2020.​29668​81
Chaki, J., & Dey, N. (2018). A beginner’s guide to image preprocessing
techniques. CRC Press.
Chan, S. S., Roth, B., Skare, M., Hernar, M., Jessen, F., Løvdal, T.,
Jakobsen, A. N., & Lerfall, J. (2020). Effect of chilling technolo-
gies on water holding properties and other quality parameters
throughout the whole value chain: From whole fish to cold-
smoked fillets of Atlantic salmon (Salmo salar). Aquaculture.
https://​doi.​org/​10.​1016/j.​aquac​ulture.​2020.​735381
Chen, L., & Opara, U. L. (2013). Approaches to analysis and mod-
eling texture in fresh and processed foods – A review. Journal
of Food Engineering, 119(3), 497–507. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​
jfood​eng.​2013.​06.​028
Cheng, J.-H., Nicolai, B., & Sun, D.-W. (2017). Hyperspectral imaging
with multivariate analysis for technological parameters predic-
tion and classification of muscle foods: A review. Meat Science,
123, 182–191. https://​doi.​org/​10.​1016/j.​meats​ci.​2016.​09.​017
Cui, D., Li, H., He, J., Wang, Q., Lu, C., Hu, H., Cheng, X., & Wang,
C. (2022). Applications of water jet cutting technology in agri-
cultural engineering: A review. Applied Sciences-Basel. https://​
doi.​org/​10.​3390/​app12​188988
Diamond, K. M., Avants, B. B., & Maga, A. M. (2021). Machine
learning-based segmentation and landmarking of 2D fish images.
Integrative and Comparative Biology, 61, E1100–E1101.
Dowgiallo, A. (2008). The effect of cutting and fish-orientation sys-
tems on the deheading yield of carp. International Journal of
Food Science and Technology, 43(9), 1688–1692. https://d​ oi.o​ rg/​
10.​1111/j.​1365-​2621.​2008.​01750.x
Dowgiallo, A., & Dutkiewicz, D. (2007). Possibilities of utilizing the
differences of fish tissues stiffness in the mechanization of cypri-
nid deheading. Journal of Food Engineering, 83(1), 111–115.
https://​doi.​org/​10.​1016/j.​jfood​eng.​2007.​01.​028
Dowlati, M., Mohtasebi, S. S., & de la Guardia, M. (2012). Application
of machine-vision techniques to fish-quality assessment. TrAC
Trends in Analytical Chemistry, 40, 168–179. https://​doi.​org/​10.​
1016/j.​trac.​2012.​07.​011
Food and Bioprocess Technology (2024) 17:828–849
Duong, L. N. K., Al-Fadhli, M., Jagtap, S., Bader, F., Martindale,
W., Swainson, M., & Paoli, A. (2020). A review of robotics
and autonomous systems in the food industry: From the supply
chains perspective. Trends in Food Science & Technology, 106,
355–364. https://​doi.​org/​10.​1016/j.​tifs.​2020.​10.​028
Einarsdottir, H., Guethmundsson, B., & Omarsson, V. (2022). Automa-
tion in the fish industry. Animal Frontiers, 12(2), 32–39. https://​
doi.​org/​10.​1093/​af/​vfac0​20
Erikson, U., Misimi, E., & Fismen, B. (2010). Bleeding of anaesthe-
tized and exhausted Atlantic salmon: Body cavity inspection and
residual blood in pre-rigor and smoked fillets as determined by
various analytical methods. Aquaculture Research, 41(4), 496–
510. https://​doi.​org/​10.​1111/j.​1365-​2109.​2009.​02338.x
FAO. (2022). The state of world fisheries and aquaculture 2022.
Towards Blue Transformation, Rome. https://​doi.​org/​10.​4060/​
cc046​1en
Gupta, K. (2020). A review on green machining techniques. Procedia
Manufacturing, 51, 1730–1736. https://d​ oi.o​ rg/1​ 0.1​ 016/j.p​ romfg.​
2020.​10.​241
Hao, Y., Yin, H., & Li, D. (2022). A novel method of fish tail fin
removal for mass estimation using computer vision. Computers
and Electronics in Agriculture. https://d​ oi.o​ rg/1​ 0.1​ 016/j.c​ ompag.​
2021.​106601
Hassoun, A., Aït-Kaddour, A., Abu-Mahfouz, A. M., Rathod, N. B.,
Bader, F., Barba, F. J., Biancolillo, A., Cropotova, J., Galanakis,
C. M., Jambrak, A. R., Lorenzo, J. M., Måge, I., Ozogul, F., &
Regenstein, J. (2022a). The fourth industrial revolution in the food
industry—Part I: Industry 4.0 technologies. Critical Reviews in
Food Science and Nutrition. https://​doi.​org/​10.​1080/​10408​398.​
2022.​20347​35
Hassoun, A., Cropotova, J., Trollman, H., Jagtap, S., Garcia-Garcia, G.,
Parra-López, C., Nirmal, N., Özogul, F., Bhat, Z., Aït-Kaddour,
A., & Bono, G. (2023). Use of industry 4.0 technologies to reduce
and valorize seafood waste and by-products: A narrative review
on current knowledge. Current Research in Food Science, 6,
100505. https://​doi.​org/​10.​1016/j.​crfs.​2023.​100505
Hassoun, A., Siddiqui, S. A., Smaoui, S., Ucak, İ, Arshad, R. N., Garcia-
Oliveira, P., Prieto, M. A., Aït-Kaddour, A., Perestrelo, R., Câmara,
J. S., & Bono, G. (2022b). Seafood processing, preservation, and
analytical techniques in the age of Industry 4.0. Applied Sciences.
https://​doi.​org/​10.​3390/​app12​031703
Hlavacova, I. M., & Geryk, V. (2017). Abrasives for water-jet cutting
of high-strength and thick hard materials. International Journal
of Advanced Manufacturing Technology, 90(5–8), 1217–1224.
https://​doi.​org/​10.​1007/​s00170-​016-​9462-y
Jayraj, P., Machavaram, R., Sahu, G., & Paradkar, V. (2019). Meas-
urement of morphometric dimensions and mechanical proper-
ties of Rohu fish for design of processing machines. Journal
of Aquatic Food Product Technology, 28(2), 150–164. https://​
doi.​org/​10.​1080/​10498​850.​2019.​15697​41
Jeebhay, M. F., Robins, T. G., Miller, M. E., Bateman, E., Smuts, M.,
Baatjies, R., & Lopata, A. L. (2008). Occupational allergy and
asthma among salt water fish processing workers. American
Journal of Industrial Medicine, 51(12), 899–910. https://​doi.​
org/​10.​1002/​ajim.​20635
Kapłonek, W., Nadolny, K., Zieliński, B., Plichta, J., Pimenov, D. Y.,
& Sharma, S. (2020). The role of observation–Measurement
methods in the surface characterization of X39Cr13 stainless-
steel cutting blades used in the fish processing industry. Mate-
rials. https://​doi.​org/​10.​3390/​ma132​45796
Kasperowicz, M. B., Chomka, G. P., & Bil, T. (2019). Determina-
tion of supply pressure during cutting fish using high-pressure
water stream taking into account the cutting place and diameter
of the water nozzle. International Journal of Food Engineer-
ing. https://​doi.​org/​10.​1515/​ijfe-​2018-​0395
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
847
Kelkar, S., Boushey, C. J., & Okos, M. (2015). A method to determine
the density of foods using X-ray imaging. Journal of Food Engi-
neering, 159, 36–41. https://​doi.​org/​10.​1016/j.​jfood​eng.​2015.​
03.​012
Khodabandehloo, K. (2022). Achieving robotic meat cutting. Animal.
Frontiers, 12(2), 7–17. https://​doi.​org/​10.​1093/​af/​vfac0​12
Kirtil, E., & Oztop, M. H. (2016). 1H Nuclear magnetic resonance
relaxometry and magnetic resonance imaging and applications
in food science and processing. Food Engineering Reviews,
8(1), 1–22. https://​doi.​org/​10.​1007/​s12393-​015-​9118-y
Kogiannou, D., Kotsiri, M., & Grigorakis, K. (2022). A method
to assess gaping in Sparidae species fillets. Aquaculture
Research, 53(2), 689–693. https://​doi.​org/​10.​1111/​are.​15590
Koulouris, A., Misailidis, N., & Petrides, D. (2021). Applications
of process and digital twin models for production simulation
and scheduling in the manufacturing of food ingredients and
products. Food and Bioproducts Processing, 126, 317–333.
https://​doi.​org/​10.​1016/j.​fbp.​2021.​01.​016
Larouche, O., Cloutier, R., & Zelditch, M. L. (2015). Head, body and
fins: Patterns of morphological integration and modularity in
fishes. Evolutionary Biology, 42(3), 296–311. https://​doi.​org/​
10.​1007/​s11692-​015-​9324-9
Li, D., Miao, Z., Peng, F., Wang, L., Hao, Y., Wang, Z., Chen, T., Li,
H., & Zheng, Y. (2021). Automatic counting methods in aqua-
culture: A review. Journal of the World Aquaculture Society,
52(2), 269–283. https://​doi.​org/​10.​1111/​jwas.​12745
Li, D., Wang, Q., Li, X., Niu, M., Wang, H., & Liu, C. (2022).
Recent advances of machine vision technology in fish clas-
sification. ICES Journal of Marine Science, 79(2), 263–284.
https://​doi.​org/​10.​1093/​icesj​ms/​fsab2​64
Lima, D. A. S., Santos, M. M. F., Duvale, R. L. F., Bezerra, T. K.
A., Araujo, I. B. D., Madruga, M. S., & da Silva, F. A. P.
(2021). Technological properties of protein hydrolysate from
the cutting byproduct of serra Spanish mackerel (Scomb-
eromorus brasiliensis). Journal of Food Science and Tech-
nology-Mysore, 58(8), 2952–2962. https://​d oi.​o rg/​1 0.​1 007/​
s13197-​020-​04797-5
Lima, D. A. S., Santos, M. M. F., Sousa, A. M. B. L., Bezerra, T. K.
A., da Silva Araújo, Í. B., Madruga, M. S., & da Silva, F. A.
P. (2022). The cutting by-product of fish filleting on the band
saw machine: Nutritional quality and technological potential.
Waste Biomass Valorization, 13(11), 4575–4584. https://​doi.​
org/​10.​1007/​s12649-​022-​01818-6
Liu, D., Zeng, X.-A., & Sun, D.-W. (2013). NIR spectroscopy and
imaging techniques for evaluation of fish quality—A review.
Applied Spectroscopy Reviews, 48(8), 609–628. https://​doi.​org/​
10.​1080/​05704​928.​2013.​775579
Liu, S., Wang, H., & Cai, Y. (2021). Research on fish slicing method
based on simulated annealing algorithm. Applied Sciences-Basel.
https://​doi.​org/​10.​3390/​app11​146503
Liu, W., Lyu, J., Wu, D., Cao, Y., Ma, Q., Lu, Y., & Zhang, X. (2022).
Cutting techniques in the fish industry: A critical review. Foods.
https://​doi.​org/​10.​3390/​foods​11203​206
Liu, X. C., Liang, Z. W., Wen, G. L., & Yuan, X. F. (2019). Waterjet
machining and research developments: A review. International
Journal of Advanced Manufacturing Technology, 102(5–8),
1257–1335. https://​doi.​org/​10.​1007/​s00170-​018-​3094-3
Lopez, M., Maia, P., Figueiredo, R., Alves, N., & Gonzalez, D. (2019).
Automated vision system for cutting fixed-weight or fixed-length
frozen fish portions. In Proceedings of the 8th International
Conference on Pattern Recognition Applications and Methods,
Prague, Czech Republic. https://​doi.​org/​10.​5220/​00074​82407​
070714
Mason, A., Romanov, D., Cordova-Lopez, L. E., Ross, S., & Korostynska,
O. (2022). Smart knife: Technological advances towards smart cutting
13
848
tools in meat industry automation. Sensor Review, 42(1), 155–163.
https://​doi.​org/​10.​1108/​sr-​09-​2021-​0315
McGeough, J. A. (2016). Cutting of food products by ice-particles in
a water-jet. In 18th CIRP Conference on Electro Physical and
Chemical Machining (ISEM XVIII), Tokyo, Japan. https://​doi.​
org/​10.​1016/j.​procir.​2016.​03.​009
Merkin, G. V., Stien, L. H., Pittman, K., & Nortvedt, R. (2013). Digi-
tal image analysis as a tool to quantify gaping and morphology
in smoked salmon slices. Aquacultural Engineering, 54, 64–71.
https://​doi.​org/​10.​1016/j.​aquae​ng.​2012.​11.​003
Mery, D., Lillo, I., Loebel, H., Riffo, V., Soto, A., Cipriano, A., & Aguilera,
J. M. (2011). Automated fish bone detection using X-ray imaging.
Journal of Food Engineering, 105(3), 485–492. https://​doi.​org/​10.​
1016/j.​jfood​eng.​2011.​03.​007
Misimi, E., Øye, E. R., Sture, Ø., & Mathiassen, J. R. (2017). Robust
classification approach for segmentation of blood defects in cod
fillets based on deep convolutional neural networks and support
vector machines and calculation of gripper vectors for robotic
processing. Computers and Electronics in Agriculture, 139,
138–152. https://​doi.​org/​10.​1016/j.​compag.​2017.​05.​021
Murthy, L. N., Phadke, G. G., Siddaiah, V., & Boraiah, R. K. (2017).
Rheological properties of washed and unwashed tilapia (Oreo-
chromis mossambicus) fish meat: Effect of sucrose and sorbitol.
Food Science and Biotechnology, 26(5), 1177–1183. https://​doi.​
org/​10.​1007/​s10068-​017-​0162-7
Nakashima, Y. (2020). Development of a hand-held magnetic reso-
nance sensor for the nondestructive quantification of fat and
lean meat of fresh tuna. Journal of Food Measurement and
Characterization, 14(6), 2947–2955. https://​doi.​org/​10.​1007/​
s11694-​020-​00539-5
Olsen, S. H., Sorensen, N. K., Larsen, R., Elvevoll, E. O., & Nilsen, H.
(2008). Impact of pre-slaughter stress on residual blood in fil-
let portions of farmed Atlantic cod (Gadus morhua) - Measured
chemically and by Visible and Near-infrared spectroscopy. Aqua-
culture, 284(1–4), 90–97. https://​doi.​org/​10.​1016/j.​aquac​ulture.​
2008.​07.​042
Ooi, C. K., Lewis, T., Nowak, B., Lyle, J., & Haddy, J. (2022). The use
of image analysis techniques for the study of muscle melanisa-
tion in sand flathead (Platycephalus bassensis)*. Environmental
Pollution. https://​doi.​org/​10.​1016/j.​envpol.​2021.​118360
Ørnholt-Johansson, G., Gudjónsdóttir, M., Nielsen, M. E., Skytte, J.
L., & Frosch, S. (2017). Analysis of the production of salmon
fillet – Prediction of production yield. Journal of Food Engineer-
ing, 204, 80–87. https://​doi.​org/​10.​1016/j.​jfood​eng.​2017.​02.​022
Prasetyo, E., Suciati, N., & Fatichah, C. (2022). Yolov4-tiny with wing
convolution layer for detecting fish body part. Computers and
Electronics in Agriculture. https://​doi.​org/​10.​1016/j.​compag.​
2022.​107023
Ren, Q.-S., Fang, K., Yang, X.-T., & Han, J.-W. (2022). Ensuring the
quality of meat in cold chain logistics: A comprehensive review.
Trends in Food Science & Technology, 119, 133–151. https://d​ oi.​
org/​10.​1016/j.​tifs.​2021.​12.​006
Riorlykhaug, E., & Egeland, O. (2019). Vision system for quality
assessment of robotic cleaning of fish processing plants using
CNN. IEEE Access, 7, 71675–71685. https://​doi.​org/​10.​1109/​
ACCESS.​2019.​29196​56
Romanov, D., Korostynska, O., Lekang, O. I., & Mason, A. (2022).
Towards human-robot collaboration in meat processing: Chal-
lenges and possibilities. Journal of Food Engineering. https://​
doi.​org/​10.​1016/j.​jfood​eng.​2022.​111117
Ross, K., & Edwards, J. (2015). Spatial variation in the mercury con-
centration of muscle myomeres in steaks of farmed Southern
Bluefin Tuna. Foods, 4(2), 254–262. https://​doi.​org/​10.​3390/​
foods​40202​54
Rustad, T., Storro, I., & Slizyte, R. (2011). Possibilities for the uti-
lisation of marine by-products. International Journal of Food
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Food and Bioprocess Technology (2024) 17:828–849
Science and Technology, 46(10), 2001–2014. https://​doi.​org/​10.​
1111/j.​1365-​2621.​2011.​02736.x
Saberioon, M., Gholizadeh, A., Cisar, P., Pautsina, A., & Urban, J.
(2017). Application of machine vision systems in aquaculture
with emphasis on fish: State-of-the-art and key issues. Reviews
in Aquaculture, 9(4), 369–387. https://d​ oi.o​ rg/1​ 0.1​ 111/r​ aq.1​ 2143
Saha, D., & Manickavasagan, A. (2021). Machine learning techniques
for analysis of hyperspectral images to determine quality of food
products: A review. Current Research in Food Science, 4, 28–44.
https://​doi.​org/​10.​1016/j.​crfs.​2021.​01.​002
Sarah, S., Anne, S., John, G., Keith, S., Philip, C., Stefan, T., & Jonathan,
D. W. (2019). Evaluation of Salmon (Salmo salar) and Rainbow
Trout (Oncorhynchus mykiss) pin bones using textural analysis
and micro X-ray computational tomography. Journal of Food Sci-
ence and Technology-Mysore, 56(7), 3313–3319. https://​doi.​org/​
10.​1007/​s13197-​019-​03803-9
Schneider, Y., Zahn, S., & Rohm, H. (2011). Ultrasonic cutting of
foods. In H. Feng, G. Barbosa-Canovas, & J. Weiss (Eds.), Ultra-
sound Technologies for Food and Bioprocessing. pp. 211–237).
Springer New York, New York, NY. https://​doi.​org/​10.​1007/​
978-1-​4419-​7472-3_9
Schuldt, S., Schneider, Y., & Rohm, H. (2018). High-speed cutting
of foods: Cutting behavior and initial cutting forces. Journal of
Food Engineering, 230, 55–62. https://​doi.​org/​10.​1016/j.​jfood​
eng.​2018.​02.​024
Secci, G., & Parisi, G. (2016). From farm to fork: Lipid oxidation
in fish products. A review. Italian Journal of Animal Science,
15(1), 124–136. https://​d oi.​o rg/​1 0.​1 080/​1 8280​5 1X.​2 015.​
11286​87
Siche, R., Vejarano, R., Aredo, V., Velasquez, L., Saldaña, E., &
Quevedo, R. (2016). Evaluation of food quality and safety with
hyperspectral imaging (HSI). Food Engineering Reviews, 8(3),
306–322. https://​doi.​org/​10.​1007/​s12393-​015-​9137-8
Simat, V., Miletic, J., Bogdanovic, T., Poljak, V., & Mladineo, I.
(2015). Role of biogenic amines in the post-mortem migration
of Anisakis pegreffii (Nematoda: Anisakidae Dujardin, 1845)
larvae into fish fillets. International Journal of Food Microbiol-
ogy, 214, 179–186. https://​doi.​org/​10.​1016/j.​ijfoo​dmicro.​2015.​
08.​008
Sivertsen, A. H., Chu, C.-K., Wang, L.-C., Godtliebsen, F., Heia, K.,
& Nilsen, H. (2009). Ridge detection with application to auto-
matic fish fillet inspection. Journal of Food Engineering, 90(3),
317–324. https://​doi.​org/​10.​1016/j.​jfood​eng.​2008.​06.​035
Sivertsen, A. H., Heia, K., Hindberg, K., & Godtliebsen, F. (2012).
Automatic nematode detection in cod fillets (Gadus morhua L.)
by hyperspectral imaging. Journal of Food Engineering, 111(4),
675–681. https://​doi.​org/​10.​1016/j.​jfood​eng.​2012.​02.​036
Sivertsen, A. H., Heia, K., Stormo, S. K., Elvevoll, E., & Nilsen, H.
(2011). Automatic nematode detection in cod fillets (Gadus
Morhua) by transillumination hyperspectral imaging. Journal of
Food Science, 76(1), S77–S83. https://​doi.​org/​10.​1111/j.​1750-​
3841.​2010.​01928.x
Skjelvareid, M. H., Heia, K., Olsen, S. H., & Stormo, S. K. (2017). Detec-
tion of blood in fish muscle by constrained spectral unmixing of
hyperspectral images. Journal of Food Engineering, 212, 252–261.
https://​doi.​org/​10.​1016/j.​jfood​eng.​2017.​05.​029
Soltani Firouz, M., & Sardari, H. (2022). Defect detection in fruit and
vegetables by using machine vision systems and image process-
ing. Food Engineering Reviews, 14(3), 353–379. https://​doi.​org/​
10.​1007/​s12393-​022-​09307-1
Song, S., Liu, Z., Huang, M., Zhu, Q., Qin, J., & Kim, M. S. (2020).
Detection of fish bones in fillets by Raman hyperspectral imaging
technology. Journal of Food Engineering. https://​doi.​org/​10.​1016/j.​
jfood​eng.​2019.​109808
Spagnoli, A., Brighenti, R., Terzano, M., & Artoni, F. (2019). Cutting
resistance of soft materials: Effects of blade inclination and friction.
Food and Bioprocess Technology (2024) 17:828–849
Theoretical and Applied Fracture Mechanics, 101, 200–206.
https://​doi.​org/​10.​1016/j.​tafmec.​2019.​02.​017
Sung, H. J., Park, M. K., & Choi, J. W. (2020). Automatic grader for
flatfishes using machine vision. International Journal of Control
Automation and Systems, 18(12), 3073–3082. https://​doi.​org/​10.​
1007/​s12555-​020-​0007-7
Urazoe, K., Kuroki, N., Maenaka, A., Tsutsumi, H., Iwabuchi, M.,
Fuchuya, K., Hirose, T., & Numa, M. (2021). Automated fish
bone detection in X-ray images with convolutional neural network
and synthetic image generation. IEEJ Transactions on Electrical
and Electronic Engineering, 16(11), 1510–1517. https://​doi.​org/​
10.​1002/​tee.​23448
Uyar, R., & Erdogdu, F. (2009). Potential use of 3-dimensional scanners
for food process modeling. Journal of Food Engineering, 93(3),
337–343. https://​doi.​org/​10.​1016/j.​jfood​eng.​2009.​01.​034
Veliyulin, E., Misimi, E., Bondo, M., Vebenstad, P. A., & Ostvik, S. O.
(2011). A simple method for weight estimation of whole herring
(Clupea harengus) using planar X-ray imaging. Journal of Food
Science, 76(3), E328–E331. https://​doi.​org/​10.​1111/j.​1750-​3841.​
2011.​02093.x
Verboven, P., Defraeye, T., Datta, A. K., & Nicolai, B. (2020). Digital
twins of food process operations: The next step for food process
models? Current Opinion in Food Science, 35, 79–87. https://​doi.​
org/​10.​1016/j.​cofs.​2020.​03.​002
Vidaček, S., & Bugge, E. (2016). Hygienic design of fish processing
equipment. In H. Lelieveld, J. Holah, & D. Gabrić (Eds.), Hand-
book of hygiene control in the food industry (2nd ed., pp. 359–365).
Woodhead Publishing, San Diego. https://​doi.​org/​10.​1016/​B978-
0-​08-​100155-​4.​00026-1
Wang, D., Zhang, M., Mujumdar, A. S., & Yu, D. (2022). Advanced
detection techniques using artificial intelligence in processing of
berries. Food Engineering Reviews, 14(1), 176–199. https://​doi.​
org/​10.​1007/​s12393-​021-​09298-5
Wang, J., & Shanmugam, D. K. (2009). Cutting meat with bone using an
ultrahigh pressure abrasive waterjet. Meat Science, 81(4), 671–677.
https://​doi.​org/​10.​1016/j.​meats​ci.​2008.​11.​010
Wang, S., Nian, R., Cao, L., Sui, J., & Lin, H. (2015). Detection of fish
bones in cod fillets by UV illumination. Journal of Food Protection,
78(7), 1414–1419. https://​doi.​org/​10.​4315/​0362-​028X.​JFP-​14-​358
Wu, D., & Sun, D.-W. (2013). Advanced applications of hyperspectral
imaging technology for food quality and safety analysis and assess-
ment: A review - Part I: Fundamentals. Innovative Food Science
and Emerging Technologies, 19, 1–14. https://​doi.​org/​10.​1016/j.​
ifset.​2013.​04.​014
Wulfkuehler, S., Stark, S., Dietz, J., Schmidt, H., Weiss, A., & Carle,
R. (2014). Effect of water jet cutting and moderate heat treatment
on quality of fresh-cut red oak leaf lettuce (Lactuca sativa L. var.
crispa). Food and Bioprocess Technology, 7(12), 3478–3492.
https://​doi.​org/​10.​1007/​s11947-​014-​1360-4
Xie, T., Li, X., Zhang, X., Hu, J., & Fang, Y. (2021). Detection of Atlantic
salmon bone residues using machine vision technology. Food Con-
trol. https://​doi.​org/​10.​1016/j.​foodc​ont.​2020.​107787
Xu, J.-L., & Sun, D.-W. (2018). Computer vision detection of salmon
muscle gaping using convolutional neural network features.
Food Analytical Methods, 11(1), 34–47. https://​doi.​org/​10.​1007/​
s12161-​017-​0957-4
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
849
Xu, W., He, Y., Li, J., Zhou, J., Xu, E., Wang, W., & Liu, D. (2023).
Robotization and intelligent digital systems in the meat cutting
industry: From the perspectives of robotic cutting, perception, and
digital development. Trends in Food Science & Technology. https://​
doi.​org/​10.​1016/j.​tifs.​2023.​03.​018
Xu, W., Wang, J., Deng, Y., Li, J., Yan, T., Zhao, S., Yang, X., Xu, E.,
Wang, W., & Liu, D. (2022). Advanced cutting techniques for solid
food: Mechanisms, applications, modeling approaches, and future
perspectives. Comprehensive Reviews in Food Science and Food
Safety, 21(2), 1568–1597. https://​doi.​org/​10.​1111/​1541-​4337.​12896
Yang, X. L., Nian, R., Lin, H., Duan, C., Sui, J. X., & Cao, L. M. (2013).
Detection of anisakid larvae in cod fillets by UV fluorescent imag-
ing based on principal component analysis and gray value analysis.
Journal of Food Protection, 76(7), 1288–1292. https://​doi.​org/​10.​
4315/​0362-​028X.​JFP-​12-​471
Yildiz, G., Palma, S., & Feng, H. (2019). Ultrasonic cutting as a new
method to produce fresh-cut red delicious and golden delicious
apples. Journal of Food Science, 84(12), 3391–3398. https://​doi.​
org/​10.​1111/​1750-​3841.​14798
Yildiz, G., Rababah, T. M., & Feng, H. (2016). Ultrasound-assisted cut-
ting of cheddar, mozzarella and Swiss cheeses – Effects on quality
attributes during storage. Innovative Food Science & Emerging
Technologies, 37, 1–9. https://​doi.​org/​10.​1016/j.​ifset.​2016.​07.​013
Yu, C., Fan, X., Hu, Z., Xia, X., Zhao, Y., Li, R., & Bai, Y. (2020).
Segmentation and measurement scheme for fish morphological
features based on Mask R-CNN. Information Processing in Agri-
culture, 7(4), 523–534. https://​doi.​org/​10.​1016/j.​inpa.​2020.​01.​002
Zhao, S., Zhang, S., Liu, J., Wang, H., Zhu, J., Li, D., & Zhao, R. (2021).
Application of machine learning in intelligent fish aquaculture: A
review. Aquaculture. https://​doi.​org/​10.​1016/j.​aquac​ulture.​2021.​
736724
Zhong, J., Zhang, F., Lu, Z., Liu, Y., & Wang, X. (2019). High-speed dis-
play-delayed planar X-ray inspection system for the fast detection
of small fishbones. Journal of Food Process Engineering. https://​
doi.​org/​10.​1111/​jfpe.​13010
Zhou, L., Zhang, C., Liu, F., Qiu, Z., & He, Y. (2019). Application of
deep learning in food: A review. Comprehensive Reviews in Food
Science and Food Safety, 18(6), 1793–1811. https://​doi.​org/​10.​
1111/​1541-​4337.​12492
Zhu, L., Spachos, P., Pensini, E., & Plataniotis, K. N. (2021). Deep learning
and machine vision for food processing: A survey. Current Research in
Food Science, 4, 233–249. https://​doi.​org/​10.​1016/j.​crfs.​2021.​03.​009
Zieliński, B., Kapłonek, W., & Nadolny, K. (2018). Regeneration of indus-
trial cutting blades made from X39Cr13 steel used in skinning pro-
cess of pleuronectidae-family flatfishes. Journal of Mechanical and
Energy Engineering, 2(4), 277–284. https://​doi.​org/​10.​30464/​jmee.​
2018.2.​4.​277
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor (e.g. a society or other partner) holds
exclusive rights to this article under a publishing agreement with the
author(s) or other rightsholder(s); author self-archiving of the accepted
manuscript version of this article is solely governed by the terms of
such publishing agreement and applicable law.
13
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

onlineservice@springernature.com