Professional Documents
Culture Documents
Survey Paper
Survey Paper
Survey Paper
net/publication/369085556
CITATIONS READS
0 1,343
4 authors:
All content following this page was uploaded by Qhubaib Shaikh on 08 March 2023.
Abstract : The increasing crime rate and number of criminals have become significant concerns for police
departments around the world. Identifying suspects for crime incidents is a time-consuming and labor-intensive
process. Face recognition technology has the potential to assist in crime prevention and criminal identification,
but it is a difficult problem in computer vision due to the high degree of variability in facial appearance.
Accuracy and speed of identification are key issues in this field. However, the robustness of the system can be
obstructed by humans who alter their facial features. Despite these challenges, face identification is preferred
for its accuracy, cost-effectiveness, and ability to stand scrutiny in court. Other methods of criminal
identification include fingerprinting, DNA matching and Eyewitness accounts, but facial recognition is more
efficient and achievable.
Through CCTV footage. The system can generate a distinctive template for each face captured in the footage
and compare it with other images stored in the dataset. Whenever a match is detected, the corresponding
image's associated details will be displayed. Implementing this system could lead to decreased crime rates and
enhanced security in the community. Nevertheless, there are apprehensions regarding privacy issues. Potential
misuse of technology and face-to-face ethical issues Validation technology.
Keywords– Criminal Identification, Criminal Tracking, Face API, Face recognition, Feature extraction, Open
CV
I. INTRODUCTION
Throughout history, no nation has been able to achieve a completely crime-free society. Crime is
inherent in human nature, and even with advances in technology, criminals have become more adept at
committing crimes in discreet ways. However, criminals have evolved and are smarter than ever in terms of
covering tracks and not leaving behind any kind of traceable evidence. Crime investigation has a very
significant role in the police system in any country. With the increased complexity of the criminal investigation
process, police officers have to bear a lot more pressure than early days.
Face recognition and detection come into play here. The face is significant for human identity and due
to its distinguishable nature, every face is unique. Facial recognition technology is a biometric technique that
provides a non-intrusive and highly accurate means of identifying criminals. Facial recognition technology
utilizes facial features to automatically detect and verify an individual's identity from images or video frames.
The potential applications of facial recognition technology are varied and extensive. They include computer
access control, airport and private surveillance, criminal identification, and security for ATM transactions. The
face recognition system is advancing towards a next-generation smart environment that enables computers to
interact more like humans. However, in the context of automated identification systems in public places, there is
no need for human observation. Automatic face recognition is a process that involves identifying and extracting
distinct facial features from an image or video frame, and then comparing them with a database of known faces
for the purpose of classification and matching. The system comprises two primary processes: detection and
identification. There are two key methods utilized in face recognition: training and evaluation. During training,
the algorithm is fed with a sample of images to train the model on the training set.
1
www.viva-technology.org/New/IJRI
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2023)
ISSN(Online): 2581-7280
Shalinda Adikari, Kaumalee Bogahawatte [2] he researchers employed clustering techniques to group crime
data into subsets based on available evidence, and then utilized Naive Bayesian classification to identify the
most likely suspects for criminal incidents. Technology used JSF (Java Server Faces) and Prime faces for the
implementation of User Interfaces. All the widgets in user interfaces such as input fields, buttons and menus
were developed using Prime faces 3.2. Prime faces built-in Ajax based on normal JSF 2.0 Ajax API is used to
update the components when they have finished processing the algorithms. To implement the database publisher
used Oracle Database 10g Express Edition.
Apoorva.P, Impana.H.C, Siri.S.L, Varshitha.M.R, and Ramesh.B [3] proposed a robust face detection
methodology suitable for real-time environments. The authors proposed a face recognition system based on
Haar Cascade, which can effectively identify faces even when the input image does not perfectly match the
images stored in the database. The proposed system has been implemented and tested.Sudha Sharma, Mayank
Bhatt, Pratyush Sharma [4] devised an investigation that is used to automatically detect the face of the person.
The ORL dataset is used for performing experiments. The algorithm has been checked on the database with
much more than 1000 images but with some false positives, it gives more than 90% accuracy. It used multiple
Algorithm like PCA with linear Discriminant Analysis which gives accuracy of 97 percent and similar
algorithms which gave average accuracy of 91 percent.
Shiva Tamrkar, Ayush Gupta [5] proposed a methodology for criminal identification in India using face
recognition, as traditional methods such as thumbprint or matching information with FIR records are not always
effective. The authors proposed three different approaches for face recognition, including the feature-based,
holistic, and hybrid methods. The feature-based approach segments local features such as the nose and eyes to
use as input for the face detection system, while the holistic approach takes the entire face as input for detection
and recognition. The hybrid approach combines both the feature-based and holistic approach. The authors
achieved 96.2% accuracy using their proposed model.
Mr. R. Prashanth Kumar, Abdul Majeed, Farhan Pasha, A Sujith [6] offered a real-time criminal identification
system based on face recognition using a fully automatic facial recognition system. The system uses Haar
feature-based cascade classifiers and OpenCV LBPH (Local Binary Pattern Histograms) algorithms for face
detection and recognition. Detecting the location of faces remains a difficult task, and many researchers have
used the Viola-Jones framework to detect the location of objects and faces in an image. Face detection
classifiers are publicly available, such as OpenCV. The authors reported achieving 95% accuracy in their
experiments.
Alireza Chevelwalla, Ajay Gurav, Sachin Desai, Prof. Sumitra Sadhukhan [8] In computer vision, face
recognition is a highly challenging task, where achieving high accuracy and speed of identification is crucial.
Countries like Germany and Australia have implemented face recognition technology at borders and for
automatic passport control. The research goal is to evaluate various face detection and recognition methods and
propose a comprehensive solution for image-based face detection and recognition with improved accuracy.
2
www.viva-technology.org/New/IJRI
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2023)
ISSN(Online): 2581-7280
Shadman Shahriar, et.al [9] provided a study on facial recognition, which is a biometric technology that maps an
individual's facial features mathematically and stores data as a face print. It utilizes machine-learning images to
generate a feature vector that maps the object with a series of numbers. This technology is used by various
organizations, including Google and Facebook, to create digital profiles for their users. The proposed project
aims to employ facial recognition technology to identify fugitive criminals with past records. According to a
report by NCRB (National Crime Records Bureau), 70% of crimes are committed by the same criminals
repeatedly. Face recognition technology can be utilized to identify criminals by capturing pictures or video
images through cameras installed in various locations. It can also be used to identify missing children. However,
the downside is that the images captured are often blurry, less clear, and difficult for the human eye to discern.
The system proposal can efficiently recognize multiple faces and is especially useful for identifying suspicious
persons due to its quick calculation time. The system creates a distinct template for each face and compares it
with other images in the dataset. When the input face matches with an existing image, the corresponding details
are displayed. Such a system can be beneficial in reducing crime and improving security measures within an
organization.
S. T. Bharathi, Dr. B. Indrani, Dr. M. Amutha Prabakar [10], The process of data mining involves extracting
meaningful patterns and relationships from large datasets. It is a valuable tool for uncovering hidden insights
and knowledge from data that may not be apparent through traditional data analysis methods. Data mining tools
and algorithms are essential in the realm of crime detection and criminal identification. Criminology is the
process of identifying the signs and behaviour of crime. Authors have a big one: the volume of multi-attribute
crime datasets leads to more complexity in the process of identifying the character of the crime. Identification of
the offender is the main process of the next procedure improving police investigations. Using data mining tools
is an effective approach to identifying criminals based on their characteristics and the nature of the crime. The
paper suggests a supervised approach to identifying a list of suspects using a similarity measure and the
K-Medoids clustering algorithm. This algorithm groups similar offenses into distinct clusters, each with its own
set of unique features. The proposed system utilizes a set of unique functions based on similarity measurement
algorithms that use distance measurements for criminal identification. It consists of two phases: training and
testing. In the training phase, the proposed system is trained with a supervised data set that includes crime
information collected from various places in Tamil Nadu, obtained through online sources. In the testing phase,
the system identifies the cluster that is closest to the test crime by utilizing the K-Medoids clustering algorithm,
and then generates a list of criminal suspects using similarity measures. After the proposed scheme was
implemented and analyzed, the results showed a high degree of accuracy. To evaluate its performance, a
comparison was made with a related K-Means clustering algorithm that utilized the same training and test
dataset.
Face recognition for The authors are utilizing They achieved 80% Authors had used PCA
criminal CCTV footage to compare accuracy using the proposed (Principal Component
identification: An images with a criminal model. Analysis), MATLAB
Implementation of database if no fingerprints are R2013b
Principal Component found at the crime scene. The
Analysis for Face third stage involves defining
Recognition [1] the system design and its
workflow.
Intelligent Criminal The system uses an explicit It could detect suspicious Authors had used Naive
Identification clustering mechanism to human behavior. Bayesian classification
System [2] segment crime data into
3
www.viva-technology.org/New/IJRI
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2023)
ISSN(Online): 2581-7280
Automated criminal The author presents a It gives quick processing The suggested face
identification by face methodology for robust face and response time. recognition system
recognition using detection in real-time utilizing Haar Cascade
open computer environments. The system can will be executed.
vision classifiers [3] effectively identify faces even
if the stored set of images of
the person in the database
differ from the input image.
Face Recognition In this paper, an investigation It has a high rate of accuracy PCA, linear
System Using is used to automatically detect i.e. it gives more than 90% Discriminant Analysis,
Machine Learning the face of the person. The accuracy. SVM, Naives Bayes
ORL dataset is used for
Algorithm [4] performing experiments.
Real-time criminal The real-time criminal This system will be able to Haar feature-based
identification system identification system utilizes a detect faces and recognize Cascade classifier as
based on face fully automated facial faces automatically in real well as OpenCV LBPH
recognition [6] recognition system that time. The author has (Local Binary Pattern
incorporates Haar achieved 95% accuracy. Histograms)
feature-based Cascade
classifier and OpenCV LBPH
(Local Binary Pattern
Histograms) Algorithms for
detecting and recognizing
faces.
CIS: An Automated This research study is based It is a distinctive identity HOG (Histogram of
Criminal on the analysis of faces, used in neutral networks as Oriented Gradients),
Identification emotions, ages and genders to input values in order to Networks Deep Learning
System [7] identify the suspects. Face analyze and compute are used in this system
recognition, emotion, age and similarities in recognition.
gender identifications are The author has achieved
accuracy of 80% after
implemented using deep training for 1 month.
learning-based CNN
approaches.
Criminal Face Facial recognition is a Using its two degree of Viola-Jones, LDA:
Recognition System complex task in the realm of Linear Discriminant
4
www.viva-technology.org/New/IJRI
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2023)
ISSN(Online): 2581-7280
[8] computer vision, requiring freedom, the system allows Analysis, EBGM:
both precision and speed for two modes of operation, one Elastic Bunch Graph
identification purposes. Facial Fitting
recognition technology has that results in very few false
been implemented by some positives and another which
countries, including Germany
and Australia, for border results in few false
control and automated negatives.
passport checks. This paper
aims to evaluate different
methods for face detection
and recognition, and propose
a comprehensive solution for
image-based face detection
and recognition with
enhanced accuracy.
Criminals and "According to a report by the It performs well Haar Cascade Classifier,
Missing Children National Crime Records OpenCV, Python
Identification Using Bureau (NCRB), the majority with both images and Facepplib
Face recognition of crimes (70%) are videos. The results displayed
And Web Scraping committed by repeat are 90%
[9] offenders. Biometric face
recognition technology accurate.
involves capturing and storing
an individual's facial features
as a unique face print. This
project suggests the use of
face recognition technology to
identify fugitive criminals
with prior records.
Implementing this system is
expected to reduce crime rates
and enhance overall security
within our community.
5
www.viva-technology.org/New/IJRI
View publication stats
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2023)
ISSN(Online): 2581-7280
IV. CONCLUSION
The Criminal Face Identification System is a technology that has the potential to revolutionize the approach
towards preventing and identifying crimes. Given the increasing rate of crime and the growing number of
offenders, it is imperative that law enforcement agencies have access to efficient and accurate methods for
identifying suspects. While there are challenges to implementing this technology, including concerns about
privacy and ethics, the benefits of the system cannot be ignored. This paper includes a synopsis analysis of a
literature study connected to face detection systems. Through its ability to expedite and enhance criminal
identification processes, the Criminal Face Identification System holds promise for reducing crime rates and
improving public safety. However, it is critical that appropriate measures are implemented to ensure the ethical
and transparent use of this technology.
ACKNOWLEDGEMENTS
We would like to express a deep sense of gratitude towards our guide Prof. Reshma Chaudhari, Computer Engineering Department for her
constant encouragement and valuable suggestions. The work that we are able to present is possible because of her timely guidance.
REFERENCES
[1] Nagnath Aherwad, Aditya Khamparia & Deep Chokshi (2021, July) “Criminal Identification System using Facial Recognition” 1st
international conference on computational research and data analytics July 2021
[2] Kaumalee Bogahawatte & Shalinda Adikari, “Intelligent Criminal Identification System” The 8th International Conference on
Computer Science & Education (ICCSE 2013) Colombo, Sri Lanka April 26-28, 2013
[3] Apoorva.P, Impana.H.C, Siri.S.L, Varshitha.M.R & Ramesh.B, “Automated Criminal Identification by Face Recognition Using
Open Computer Vision” Proceedings of the Third International Conference on Computing Methodologies and Communication
IEEE Xplore Part Number: CFP19K25-ART; ISBN: 978-1-5386-7808-4 (ICCMC 2019)
[4] Sudha Sharma, Mayank Bhatt & Pratyush Sharma, “Face Recognition System Using Machine Learning Algorithm” Proceedings of
the Fifth International Conference on Communication and Electronics Systems IEEE Conference Record #48766; IEEE Xplore
ISBN: 978-1-7281-5371-1 (ICCES 2020)
[5] Shiva Tamrkar, Ayush Gupta, “Criminal Face Detection System Using Python”, © July 2020| IJIRT | Volume 7 Issue 2 | ISSN:
2349-6002
[6] Mr.R. Prashanth Kumar, Abdul Majeed, Farhan Pasha, A Sujith, “REAL-TIME CRIMINAL IDENTIFICATION SYSTEM BASED
ON FACE RECOGNITION” V26, Num 05, May/2020| Advanced Science Letters | E-ISSN:1936-7317
[7] Kavushica Rasanayagam, Kumarasiri S.D.D.C, Tharuka, “CIS: An Automated Criminal Identification System”, IEEE 2018.
[8] Alireza Chevelwalla, Ajay Gurav, Sachin Desai, Prof. Sumitra Sadhukhan, “Criminal Face Recognition System” International
Journal of Engineering Research & Technology (IJERT), Vol. 4 Issue 03, March-2015
[9] S. Aryyappan, Dr. S.Matilda, “ Criminals And Missing Children Identification Using Face recognition And Web Scraping” IEEE
ICSCAN 2020
[10] S. T. Bharathi, Dr. B. Indrani, Dr. M. Amutha Prabakar, “A Supervised Learning Approach for Criminal Identification Using
Similarity Measures and K-Medoids Clustering”, International Conference on Intelligent Computing, Instrumentation and Control
Technologies (ICICICT) 2017.
[11] Samit Shirsat, Aakash Naik, Darshan Tamse, “Proposed System for Criminal Detection and Recognition on CCTV Data Using
Cloud and Machine Learning”, 2019 International Conference on Vision Towards Emerging Trends in Communication and
Networking (ViTECoN)
[12] Shalince Dominic, Mahesh Mohan, Aparna C, Ajeesh M S, Aswin S Nath, Anil Antony, “A REVIEW OF FACE DETECTION
SYSTEM”, International Conference on Electrical, Electronics, and Optimization Techniques (ICEEOT) – 2016
[13] V. Q. Nhat, S.-H. Kim, J. H. Yang and G. Lee, “Real-time Face Tracking with Instability using a Feature-based Adaptive Model”,
Intrenational Journal of Control, Automatin and Systems, vol. 13, no. 3, pp. 725-732, 2015.
[14] M. Prathamesh, R. Vedant, M. Suraj, K. Avinash and S.V.Wankhade, "Criminal Tracking System using CCTV," Imperial Journal of
Interdisciplinary Research, vol. II, no. 7, pp. 206-208, 2016.
[15] G. Divya and J. Arunkant, "FACE DETECTION AND TRACKING AT DIFFERENT ANGLES IN VIDEO," ARPN Journal of
Engineering and Applied Sciences, vol. 10, no. 17, pp. 7678-7683, 2015.
6
www.viva-technology.org/New/IJRI