Professional Documents
Culture Documents
Results and Discussion
Results and Discussion
Results and Discussion
This chapter presents the data and findings gathered through the research
instruments used in the study. The data were analyzed and statistically treated in order to
evaluate the interpretations of data that will lead to the conclusions and recommendations
of the study.
1. The Stages Undertaken in the Development of the Computer Vision System for
The first phase in the development of the machine learning algorithm for meat
quality standard is the Requirements definition. This phase involves identifying the
problems and gathering necessary information and documents needed to develop the
device as well as defining its functional requirements. A thorough literature search was
device that will evaluate and classify pork meat based on quality. Selected meat
inspectors, who are identified as the respondents that will mainly benefit from the
development of the device were interviewed as well to give some ideas about meat
quality inspection.
After gathering sufficient information and considering the target users of the
device, the researcher moved to the second phase of the development which is the
Planning. The researcher has identified the scope of the problem as well as solutions to
them. After the interview with the meat inspector, the problem on the time-consuming
and subjective process of visual meat inspection surfaced which somehow cause
inefficiency when accurate and instantaneous results are needed. The researcher then
thought of, based on literature, the development of a system that will shorten the time
The third phase is Designing. Based on the information gathered from the
literature search and the interview conducted with the meat inspectors, a block diagram
was created.
INPUT UNIT CONTROL AND DISPLAY UNIT
(Raspberry pi PROCESSING UNIT (LCD)
camera) (Raspberry pi)
Figure 3. Block Diagram for the Computer Vision System for Meat Quality
Evaluation
The design in Figure 3 has helped in specifying hardware and software
requirements. The image of a meat sample is captured using a Picamera. The image
captured is then processed by the meat quality classifier software installed in the
The prototype interface of the device was designed as shown in Figure 4. The
prototype interface has two buttons: Capture and Clear. The Capture button takes the
image of the meat sample and triggers the evaluation and classification process. On the
other hand, the Clear button, erases previous classification readings and readies the
camera for another image capture. The white box frame on the left of the interface shows
the captured image of the meat sample while the classification label is shown on the
Figures 3 and 4, all requirements are translated into interfaces and codes. The computer
vision system for meat quality evaluation software was created using Python as the main
as the machine learning technique which enhanced the capability of the computer vision
system. The scripts to retrain the network using the meat sample datasets and all the
numerical calculations on these datasets were provided by Tensorflow. The models and
The fifth phase is the Release. This is where the newly developed device was
The testing included comparing the pH values of pork meat with the color-
texture-exudation readings of the system. Each samples were scanned three (3) times and
The device was also tested in three (3) public markets in Cavite. These are Silang
Public Market, Tagaytay City Market and Kadiwa Market, Dasmariñas City. The
researcher scanned the pork meat on different times of the day: morning (5:00am -
6:00am), noon (12:00nn -1:00pm) and afternoon (4:00pm). The results were recorded.
The respondents were then presented with the device and their comments and
The sixth phase is Track and Monitor where the device is re-released to the
respondents for use. After this phase, the respondents were asked to evaluate the device
The evaluation of the Meat Inspectors and IT Practitioners are presented in the
tables below. The developed device was evaluated in terms of Functional Suitability
(Table 1), Performance Efficiency (Table 2), Usability (Table 3), Reliability (Table 4),
Table 1
Weighted Mean and Verbal Interpretation of Meat Inspectors and IT Practitioners’
Evaluation on the Computer Vision System for Meat Quality Evaluation in terms of
Functional Suitability
Indicators WM VI WM VI TOTAL VI
WM
1) Completeness- 4.57 Highly 4.73 Highly 4.65 Highly
refers to the degree to Acceptable Acceptable Acceptable
which the set of
functions covers all the
specified tasks and user
objectives of the
device.
2) Correctness-refers to 4.07 Acceptable 4.30 Acceptable 4.19 Acceptable
the degree to which a
product or system
provides the correct
results with the needed
degree of precision of
the device.
3) Appropriateness- 4.32 Acceptable 4.53 Highly 4.43 Acceptable
refers to the degree to Acceptable
which the functions
facilitate the
accomplishment of
specified tasks and
objectives of the
device.
OVERALL 4.32 Acceptable 4.52 Highly 4.42 Acceptable
Acceptable
Legend: WM-Weighted Mean; VI-Verbal Interpretation
and Appropriateness, with a weighted mean of 4.32 and 4.52 respectively, having a
Acceptable and an overall weighted mean of 4.65 with a verbal interpretation of Highly
weighted mean of 4.07 and 4.30 with a verbal interpretation of Acceptable and an overall
has a weighted mean of 4.32 and 4.53 respectively, and has a verbal interpretation of
Acceptable and Highly Acceptable and an overall weighted mean of 4.43 with a verbal
interpretation of Acceptable. This shows that according to the two groups of respondents,
the device was able to provide effectivity in terms of Functional Suitability. The
evaluation result was consistent with the result of the research of Sun et al. (2018) in
which it was stated that the one area where the use of computer vision system has spread
rapidly is in the inspection of food products. The quality attributes of meat products have
been traditionally evaluated by trained personnel, but today have mostly been replaced by
automatic inspection systems based on computer vision and image analysis. It must also
be noted that development in deep learning and convolutional neural network are very
efficient for food classification and recognition. Deep learning learns the image features
and extracts contextual details and global features that will help in reducing the error
Behavior, Resource Utilization, and Capacity. Time Behavior has an evaluation of 3.96
from Meat Inspectors with a verbal interpretation of Acceptable and the evaluation of IT
Overall, Time Behavior has a weighted mean of 4.02 with a verbal interpretation of
Acceptable.
Resource Utilization has an evaluation of 4.00 from Meat Inspectors and has a
mean of 4.20 with a verbal interpretation of Acceptable. Overall, Resource Utilization got
Table 2
Weighted Mean and Verbal Interpretation of Meat Inspectors and IT Practitioners’
Evaluation on the Computer Vision System for Meat Quality Evaluation in terms of
Performance Efficiency
Indicators WM VI WM VI TOTAL VI
WM
1) Time behavior- 3.96 Acceptable 4.07 Acceptable 4.02 Acceptable
refers to the degree to
which the response and
processing times and
throughput rates of the
device, when
performing its
functions, meet
requirements.
2) Resource utilization- 4.00 Acceptable 4.20 Acceptable 4.10 Acceptable
refers to the degree to
which the amounts and
types of resources used
by the device, when
performing its
functions, meet
requirements.
3) Capacity-refers to 4.07 Acceptable 4.40 Acceptable 4.24 Acceptable
the degree to which the
maximum limits of the
device parameter meet
requirements.
OVERALL 4.01 Acceptable 4.22 Acceptable 4.12 Acceptable
Capacity has a weighted mean of 4.07 from the evaluation of Meat Inspectors
Overall, the Performance Efficiency criterion of the device from the Meat
Inspectors and IT Practitioners has an overall weighted mean of 4.01 and 4.22
respectively, with a verbal interpretation of Acceptable. This shows that according to the
two groups of respondents, the device was able to provide effectivity in terms of
process and technical evaluation of using electronic media, computer vision has the
C. Barbon et al., 2017, Sanaeifar, Bakhshipour, & de la Guardia, 2016), thus, the
Meat Inspectors with a verbal interpretation of Highly Acceptable and the evaluation of
Acceptable. Overall, Operability has a weighted mean of 4.62 with a verbal interpretation
of Highly Acceptable. User Error Protection has an evaluation of 4.29 from Meat
Acceptable. Overall, User Error Protection has a weighted mean of 4.41 with a verbal
Acceptable. Overall, User Interface Aesthetics has a weighted mean of 4.58 with a verbal
interpretation of Highly Acceptable. Accessibility has an evaluation of 4.36 from Meat
interpretation of Acceptable.
Generally, the Usability criterion of the device from the Meat Inspectors and IT
Practitioners has an overall weighted mean of 4.46 and 4.66 respectively, with a verbal
interpretation of Acceptable and Highly Acceptable. This shows that according to the two
groups of respondents, the device was able to provide effectivity in terms of Usability.
Table 4
Weighted Mean and Verbal Interpretation of Meat Inspectors and IT Practitioners’
Evaluation on the Computer Vision System for Meat Quality Evaluation in terms of
Reliability
Indicators WM VI WM VI TOTAL VI
WM
1) Maturity-refers to 3.96 Acceptable 4.10 Acceptable 4.03 Acceptable
the degree to which the
device meets needs for
reliability under normal
operation.
2) Availability-refers to 3.93 Acceptable 4.00 Acceptable 3.96 Acceptable
the degree to which the
device is operational
and accessible when
required for use.
3) Fault tolerance- 4.04 Acceptable 4.03 Acceptable 4.04 Acceptable
refers to the degree to
which the device
operates as intended
despite the presence of
hardware or software
faults.
4) Recoverability-refers 3.96 Acceptable 4.13 Acceptable 4.05 Acceptable
to the degree to which,
in the event of an
interruption or a
failure, the device can
recover the data
directly affected and re-
establish the desired
state of the system.
OVERALL 3.97 Acceptable 4.07 Acceptable 4.02 Acceptable
Tolerance, and Recoverability. Maturity has an evaluation of 3.96 from Meat Inspectors
weighted mean of 4.10 with a verbal interpretation of Acceptable. Overall, Maturity has a
evaluation of 3.93 from Meat Inspectors with a verbal interpretation of Acceptable and
the evaluation of IT Practitioners has weighted mean of 4.00 with a verbal interpretation
Overall, Fault Tolerance has a weighted mean of 4.04 with a verbal interpretation of
Acceptable. Recoverability has an evaluation of 3.96 from Meat Inspectors with a verbal
Generally, the Reliability criterion of the device from the Meat Inspectors and IT
Practitioners has an overall weighted mean of 3.97 and 4.07 respectively, with a verbal
interpretation of Acceptable. This shows that according to the two groups of respondents,
Table 5
Weighted Mean and Verbal Interpretation of Meat Inspectors and IT Practitioners’
Evaluation on the Computer Vision System for Meat Quality Evaluation in terms of
Portability
Indicators WM VI WM VI TOTAL VI
WM
1) Adaptability- 4.14 Acceptabl 4.33 Acceptabl 4.24 Acceptable
refers to the degree e e
to which the device
can effectively and
efficiently be
adapted for
different or
evolving hardware,
software or other
operational or
usage
environments.
2) Installability- 4.21 Acceptabl 4.33 Acceptabl 4.27 Acceptable
refers to the degree e e
of Acceptableness
and efficiency with
which the device
can be successfully
installed and/or
uninstalled in a
specified
environment.
3) Replicability- 4.14 Acceptabl 4.43 Acceptabl 4.29 Acceptable
refers to the degree e e
to which the device
can replace another
specified software
product for the
same purpose in
the same
environment.
OVERALL 4.17 Acceptabl 4.37 Acceptabl 4.27 Acceptable
e e
Legend: WM-Weighted Mean; VI-Verbal Interpretation
of 4.21 from Meat Inspectors with a verbal interpretation of Acceptable and the
Acceptable.
Overall, the Portability criterion of the System from the Meat Inspectors and IT
Practitioners has an overall weighted mean of 4.17 and 4.37 respectively, with a verbal
interpretation of Acceptable. This shows that according to the two groups of respondents,
Table 6
Overall Weighted Mean and Verbal Interpretation of the Respondents on the
Computer Vision System for Meat Quality Evaluation based on Functional
Suitability, Performance Efficiency, Usability, Reliability, and Portability
Indicators WM VI WM VI
Table 6 shows the overall general weighted mean of the evaluation of the
respondents on the device per criteria. Based on the criteria, the general weighted mean
of the Meat Inspectors was 4.19 with the verbal interpretation of Acceptable. For IT
Practitioners, the general weighted mean was 4.37 with a verbal interpretation of
Acceptable. This shows that, according to the two groups of respondents, the device is
Reliability, and Portability. With recent advances in hardware and software, computer
vision system has been allowed to become a technology even more cost-effective, more
consistent, more rapid, and more accurate than ever before (Sun et al., 2017).
Usability, Reliability, and Portability, the z-test is applied. The result of the application of
Table 7
Summary of the z-test Results of the Two Groups of Respondents for Functional
Suitability, Performance Efficiency, Usability, Reliability, and Portability
Criteria z value
test. The computed value of z=1.04, 0.05 level of significance with a critical value of 1.96
The result of the non-rejection of the null hypothesis indicates the equality of
evaluation between the two groups of respondents which proves that the device meets the
test. The computed value of z=1.04, 0.05 level of significance with a critical value of 1.96
The result of the non-rejection of the null hypothesis indicates the equality of
evaluation between the two groups of respondents which proves that the device meets the
Inspectors and IT Practitioners of the device in terms of Usability using z-test. The
computed value of z=0.26, 0.05 level of significance with a critical value of 1.96 accepts
The result of the non-rejection of the null hypothesis indicates the equality of
evaluation between the two groups of respondents which proves that the device meets the
3.4 Reliability
Inspectors and IT Practitioners of the device in terms of Reliability using z-test. The
computed value of z=0.60, 0.05 level of significance with a critical value of 1.96 accepts
The result of the non-rejection of the null hypothesis indicates the equality of
evaluation between the two groups of respondents which proves that the device meets the
3.5 Portability
Inspectors and IT Practitioners in the device in terms of Portability using z-test. The
computed value of z=1.39, 0.05 level of significance with a critical value of 1.96 accepts
evaluation between the two groups of respondents which proves that device meets the
It is important to note that there are several limitations to this study that should be
kept in mind when interpreting the findings. First, the transfer learning approach used by
the study builds on the information that already exists in pre-trained models. Second,
models using greater amounts of training data take longer to create, thus using a
relatively modest amount of training data is advised. As a result, these models may be
This chapter discusses the summary of findings from the gathered and analyzed
data, the conclusions drawn from the findings and recommendations offered by the
Summary of Findings
1. Stages undertaken in the development of the computer vision system for meat
The researcher used the Agile model in the development of the computer vision
system for meat quality evaluation which are further broken down into the following
phases: Requirements, Planning, Designing, Development, Release, and Track and
Monitor.
2.3 Usability
2.4 Reliability
2.5 Portability
vision has the potential to provide even better evaluations of meat quality. The meat
industry in general, may use the findings of this study as a framework for developing
plans, guidelines and necessary steps to make sure that the meat available in the market is
of quality.
Conclusions
Based from the findings of the study, there is no difference between the
Chapter 1.
1. The stages undertaken in the development of the computer vision system for meat
quality evaluation using the Agile model are: Requirements, Planning, Designing,
2. The meat inspectors and IT practitioners evaluated the computer vision system for
meat quality evaluation based on a software quality assessment tool, ISO 25010, in terms
developed device. This proved that the two groups of respondents agree that the
application is “Acceptable” in terms of Functional Suitability, Performance Efficiency,
4. With the recent advancements in algorithms and computer hardware, computer vision
Recommendations
Based from the findings and conclusions of the study, the following
1. The researcher recommends that the NMIS implement the usage of the device for
visual inspection.
2. For future studies, the researcher recommends other comparative tests aside from pH
test.