Lab Register Management System Using Facial Recognition 1
IV/IV B. TECH 2ND SEMESTER PROJECT WORK (B20CS4201)
1st REVIEW
Batch No : E02
Student Registration Number
Sri Ranganadh Sunkara 20B91A05T6
Sathvik Tadi 20B91A05T8
Abhiram Yedurada 20B91A05W7
Seetha Ram Yelisetti 20B91A05W9
Under the Guidance of
Dr. N. K. Kameswara Rao
Associate Professor
SAGI RAMA KRISHNAM RAJU ENGINEERING COLLEGE (A)
(Affiliated to JNTUK, Kakinada)
SRKR MARG, CHINNA AMIRAM, PIN:534204
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
IV/IV B. TECH 2ND SEMESTER PROJECT WORK (B20CS4201)
1st REVIEW
Lab Register Management System Using Facial Recognition
Batch No : E02
Sri SEETHARAM YELISETTI 20B91A05W9
Under the Guidance of
Dr. N. K. Kameswara Rao
Associate Professor
SAGI RAMA KRISHNAM RAJU ENGINEERING COLLEGE (A)
(Affiliated to JNTUK, Kakinada)
SRKR MARG, CHINNA AMIRAM, PIN:534204
TABLE OF CONTENTS
Problem statement
Abstract
Literature survey
Methodology
Technologies used
Queries
ABSTRACT
To enhance and overcome challenges inherent in traditional lab-register
management systems, our approach involves leveraging the capabilities
of deep
learning techniques. Specifically, we employ these techniques to identify
faces, allowing for a more automated and efficient means of maintaining
the lab register. The register itself contains crucial data, including student
registration numbers, lab subject details, as well as in-time and out-time
records.
Our incorporation of deep learning techniques in lab
register management not only addresses the
challenges of traditional systems but also opens up
new possibilities for data analysis and informed
decision-making, ultimately enhancing the overall
efficiency and effectiveness of lab operations
LITERATURE SURVEY 5
• TITLE:DeepFace: Closing the Gap to Human-Level Performance in Face Verification
(15)
• Authors: Yaniv Taigman, Ming Yang, Marc'Aurelio Ranzato, Lior Wolf
• Year of published : 2015
• Algorithm :convolutional neural networks (87%)
• Dataset used: YOUTUBE DB
• INSIGHTS
• Introduces a deep learning model for face verification.
• Utilizes a vast dataset of four million facial images for training the model.
• Employs a hierarchical representation learned from the dataset.
• Model is designed to be invariant to variations in pose and illumination.
• Significantly enhances face verification accuracy.
• Aims to close the performance gap between computer-based systems and human-level accuracy
• Highlights the efficacy of deep neural networks, particularly convolutional neural networks (CNNs), in
advancing facial recognition capabilities.
LITERATURE SURVEY 6
• DeepID3: Face Recognition with Very Deep Neural Networks (16)
• Authors: Yi Sun, Xiaogang Wang, Xiaoou Tang
• Year of published : 2015
• Algorithm :Deep neural networks (92%)
• Data used: CASIA WEB FACE
• INSIGHTS
• The paper aims to improve face recognition through the development of the DeepID3 model.
• Utilizes very deep neural networks to learn hierarchical representations of facial features.
• DeepID3 focuses on effective feature learning to capture intricate patterns and variations in facial
images.
• Trains the model on large-scale datasets to enhance its ability to generalize across diverse facial
characteristics.
• Demonstrates improved face recognition performance compared to existing methods
• The paper discusses the potential applications of DeepID3 in various scenarios, including surveillance
systems, access control, and other areas where facial recognition is essential
LITERATURE SURVEY 7
• TITLE:"SphereFace: Deep Hypersphere Embedding for Face Recognition“ (17)
• Authors: Weiyang Liu, Yandong Wen, Zhiding Yu, Ming Li
• Year of published : 2017
• Algorithm :Resnet64 (82%)
• Dataset used: Facebook profile DB
• INSIGHTS
• Enhance face recognition accuracy through deep learning techniques.
• Utilizes a hypersphere for embedding face features.
• Introduces an angular margin penalty to enforce constraints on angular differences between features of
different identities.
• Implements the Angular Softmax loss function to encourage clear separation between face feature
representations.
• Utilizes a deep neural network for hierarchical learning of face features optimized for hypersphere
mapping.
• Emphasizes the significance of hypersphere embedding in simplifying decision boundaries and improving
generalization in face recognition systems.
LITERATURE SURVEY 8
• TITLE:VGGFace2: A dataset for recognising faces across pose and age (18)
• Authors: Sun Qiong Cao, Li Shen, Weidi Xie, Omkar M. Parkhi, Andrew Zisserman
• Year of published : 2018
• Algorithm :Deep CNN (87%)
• Dataset used:vggFace2
• INSIGHTS
• VGGFace2 is constructed as a large-scale face recognition dataset with a diverse set of subjects,
addressing challenges related to variations in pose, age, and other factors
• The dataset comprises millions of images collected from the internet, ensuring a broad range of facial
variations and encompassing diverse demographics.
• VGGFace2 provides rich annotations, including identity labels, age information, and pose variations,
facilitating research on recognizing faces across different conditions.
• he dataset intentionally includes images with extreme pose variations to evaluate and improve the
robustness of face recognition algorithms under challenging conditions.
• VGGFace2 serves as a benchmark for evaluating and comparing the performance of face recognition
algorithms, especially concerning their ability to handle variations in pose and age.
LITERATURE SURVEY 9
• TITLE:ArcFace: Additive Angular Margin Loss for Deep Face Recognition (19)
• Authors: Jiankang Deng, Jia Guo, Niannan Xue, Stefean
• Year of published : 2018
• Algorithm :ANGULAR MARGIN CNN (87%)
• Dataset used: ms celeb
• INSIGHTS
• ArcFace proposes an additive angular margin loss function that enforces a margin between classes in the
feature space, enhancing the discriminative power of learned features.
• The model embeds angular information into the learning process, ensuring that the feature
representations maintain a distinct separation between different classes
• ArcFace aims to achieve a sphere-like decision boundary in the feature space, promoting better
generalization and robustness in face recognition tasks.
• The paper demonstrates that ArcFace outperforms traditional softmax loss in terms of face recognition
accuracy, particularly in scenarios with a large number of classes.
• The proposed ArcFace method achieves state-of-the-art performance on various face recognition
benchmarks, showcasing its effectiveness in learning discriminative features
LITERATURE SURVEY 10
• TITLE:Facial Expression Recognition using Facial Landmarks: A novel approach (20)
• Author: Rohith Raj S, Pratiba D, Ramakanth Kumar P
• Year of publish :2020
• Algorithm Used: SVC,CNN
• INSIGHT
• The study focuses on classifying facial expressions into anger, contempt, disgust, fear, happiness,
sadness, and surprise.
• Applications include psychology, entertainment, and medical research
• Proposed methodology involves CLAHE, facial landmark prediction, and SVM for emotion recognition
• FER based on facial landmarks, achieving comparable performance to CNN-based methods using a
Support Vector Classifier.
• Achieved an average accuracy of 89% using the CK+ dataset
PROBLEM 11
STATEMENT
• The lab register, though rooted in tradition, remains a
cornerstone of efficient lab management, balancing simplicity
with the evolving needs of a dynamic research environment
Solution
• Using Deep Learning to automate register
management using facial recognition
METHODOLOGY
ARCHITECTURE
Requirement Analysis Data Collection Model Selection
Model Training Database integration UI Development
Testing & Evaluation Deployment
Database
StudentID Name Year Section ID Subject Date Subje Shift
Code ct
20B91A05 Seetharam 4 E Nam
w9 e
20B91A05 Sathvik 4 E 1 B20HS410 2024-01- FLAT Forenoo
w7 1 29 n
2 B20HS410 2024-01- FLAT Afternoo
1 29 n
Attendanc StudentID ScheduleI In Time Out Time
eID D
20B91A05T 20B91A05T 1 9:01:43 12:05:59
6_1 6
20B91A05T 20B91A05T 1 9:02:57 12:01:12
8_1 8
Deep
Learning ,
OpenCV
ResNet50 Frontend : HTML, CSS, Java Script,
Algorithm Tkinter
Software Server side : Python
Requirements Flask
Database : SQLite3
Ram : 8 GB or More
Hardware
SDD (recommended)
Requirements
Webcam 720p
TECHONOLOGIES USED
• DEEP LEARNING,OPEN CV
• RESNET50 ALGOTITHM
• SOFTWARE REQURIMENT
Front end : HTML, CSS, Java Script, Tinkter
Server side : Python Flask
Database : SQLite3
• HARDWARE REQURIMENT
Ram : 8 GB +
SDD (recommended)
Webcam 720p
REFERENCES
1. M. T. H. Fuad et al "Recent Advances in Deep Learning Techniques for Face
Recognition"- IEEE Access(2020)
2. M. Arsenovic, S. Sladojevic, A. Anderla and D. Stefanovic "FaceTime - Deep learning
based face recognition attendance system"- IEEE Access(2017)
3. M. H. Robin, M. M. Ur Rahman, A. M. Taief and Q. Nahar Eity "Improvement of Face
and Eye Detection Performance by Using Multi-task Cascaded Convolutional
Networks"- IEEE(2020)
4. A. Singh "Face Detection using Deep Recurrent Learning and SMQT Technique"-
IEEE(2020)
5. T. D. Tithy, S. Chakraborty, R. Islam and A. Aziz "A Deep Learning based Approach for
Real Time Face Recognition System"-IEEE(2021)
6. S. D. Lin and P. E. Linares Otoya “Pose-Invariant Face Recognition Using Ensemble
Learning and Local Feature Descriptors”-IEEE ACCESS(2023)
7. Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun “Deep Residual Learning for
Image Recognition”-IEEE(2016)
8. Rohith Raj S, Pratiba D, Ramakanth Kumar P “Facial Expression Recognition using
Facial Landmarks: A novel approach”-IEEE(2020)
10)Florian Schroff, Dmitry Kalenichenko, and James Philbin “FaceNet: A Unified Embedding
for Face Recognition and Clustering”-IEEE(2015)
11). S. D. Lin and P. E. Linares Otoya “Pose-Invariant Face Recognition Using Ensemble
Learning and Local Feature Descriptors”-IEEE(2023)
12). Carlijn Meijerink” Facial Landmark Detection Under Challenging Conditions”-IEEE(2018)
13). Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun” Deep Residual Learning for
Image Recognition – IEEE(2016)
14). Rohith Raj S, Pratiba D, Ramakanth Kumar P” Facial Expression Recognition using
Facial Landmarks: A novel approach “-IEEE(2020)
15). Yaniv Taigman, Ming Yang, Marc'Aurelio Ranzato, Lior Wolf” DeepFace: Closing the
Gap to Human-Level Performance in Face Verification-IEEE(2015)
16). Yi Sun, Xiaogang Wang, Xiaoou Tang“DeepID3: Face Recognition with Very Deep
Neural Networks”-IEEE(2015)
17). Weiyang Liu, Yandong Wen, Zhiding Yu, Ming Li "SphereFace: Deep Hypersphere
Embedding for Face Recognition" IEEE(2017)
19). Sun Qiong Cao, Li Shen, Weidi Xie, Omkar M. Parkhi, Andrew Zisserman”
VGGFace2: A dataset for recognising faces across pose and age”-IEEE(2018)
20). Jiankang Deng, Jia Guo, Niannan Xue, Stefean” Additive Angular Margin
Loss for Deep Face Recognition-IEEE(2015)
ANY QUERIES
THANK YOU
Yelisetti seetharam
