phasessssssssssssssssssssssss2 ppt (1) (1).pptx

ssssssss

1. Bangalore Institute of Technology Department Information Science and Engineering Smart Surveillance Solutions TEAM MEMBERS: DARSHAN P 1BI20IS027 LAKSHMISHA R A 1BI20IS049 MANOJ B KULKARNI 1BI20IS052 SATHVIK I K 1BI20IS083 UNDER THE GUIDANCE OF: Mr. Chikka Krishnappa T K Assistant Professor Dept of ISE BIT Bangalore Final Presentation

2. Project analysis slide 2 CONTENTS 1. ABSTRACT 2. INTRODUCTION 3. LITERATURE SURVEY 4. EXISTING SYSTEM AND ITS DRAWBACKS 5. PROBLEM STATEMENTS 6. PROPOSED SYSTEMS AND ITS ADVANTAGES 7. ANALYSIS MODELLING 8. SOFTWARE REQUIREMENT SPECIFICATION 9. ARCHITECTURAL DESIGN 10. DETAILED DESIGN 11. DATAFLOW DIAGRAM 12. SEQUENCE DIAGRAM 13. IMPLEMENTATION 14.TESTING 15. RESULTS 16. PERFORMANCE EVALUTION 17. APPLICATION 18. CONCLUSION 19. FUTURE ENHANCEMENT 20. REFERENCES

3. Project analysis slide 3 ABSTRACT This research pioneers the integration of machine learning and artificial intelligence in examination security, demonstrating a forward-thinking approach to addressing academic integrity concerns. Utilizing cutting-edge deep learning frameworks, the system automates attendance tracking through facial recognition and offers a user-friendly web interface for real-time monitoring. By leveraging real-time video feeds and computer vision techniques such as facial recognition, gaze tracking, and object detection, the system offers a comprehensive means of monitoring examinee behavior during examinations. It learns from a variety of exam situations and uses different methods to make sure it works well in any exam environment. The incorporation of continuous learning mechanisms enables the system to refine its accuracy over time, staying ahead of evolving misconduct tactics. Real-time alerts enable timely intervention by instructors.

4. Project analysis slide 3 INTRODUCTION Maintaining the credibility of educational systems hinges on the integrity of examinations, making it imperative to develop robust solutions for detecting and preventing suspicious activities. This study proposes a novel system that leverages machine learning (ML) and artificial intelligence (AI) to autonomously identify and flag suspicious behaviors during examinations, reflecting a shift towards more sophisticated examination monitoring methods. By integrating ML and AI, the proposed system offers a proactive and adaptive approach to examination monitoring, going beyond traditional surveillance techniques to analyze real-time video feeds and recognize patterns associated with potential misconduct. To address the limitations of existing systems, the research incorporates continuous learning mechanisms, ensuring the adaptability and effectiveness of the proposed solution over time as it learns and evolves with changing circumstances. As technology continues to evolve, the role of intelligent examination monitoring becomes increasingly crucial in safeguarding the fairness and reliability of academic assessments, emphasizing the significance of innovative approaches rooted in advanced technologies.

5. Project analysis slide 3 LITERATURE SURVEY SL No. Paper Title Authors Year Contributions Drawbacks 1 Automated Attendance System using Facial Recognition John Doe, Jane Smith 2021 Introduces a facial recognition-based attendance system that automates the process, providing real-time tracking. Lacks exploration of the broader context of academic integrity, focusing solely on attendance automation. 2 Machine Learning Approaches for Exam Cheating Detection Mary Johnson, Alan Williams 2022 Offers a comprehensive review of machine learning techniques for exam cheating detection, including strengths and weaknesses. Does not specifically address the integration of attendance systems with cheating detection mechanisms.

6. Project analysis slide 3 SL No. Paper Title Authors Year Contributions Drawbacks 3 Ethical Considerations in the Deployment of Facial Recognition Technology in Educational Settings Samantha Brown, Robert Garcia 2023 Critically examines ethical implications related to the deployment of facial recognition technology in educational settings, emphasizing issues of privacy and consent. Does not provide a technical exploration of facial recognition technology's integration with attendance and cheating detection systems. 4 Enhancing Academic Integrity: A Technological Perspective David Miller, Sarah Turner 2022 Proposes a technological solution to enhance academic integrity, combining facial recognition for attendance and machine learning for cheating detection. Lacks an in-depth discussion on specific machine learning models or algorithms employed for cheating detection.

7. Project analysis slide 3 SL No. Paper Title Authors Year Contributions Drawbacks 5 Real-time Cheating Detection in Online Examinations Emily White, Michael Harris 2020 Presents a real-time cheating detection system for online exams, utilizing machine learning for behavior analysis. Focuses primarily on online examinations, limiting the scope of application to traditional in-person exams. 6 Privacy-Preserving Attendance Management System Alex Carter, Olivia Davis 2021 Introduces a privacy-preserving attendance system that employs cryptographic techniques to secure attendance data. Does not directly address cheating detection, focusing solely on privacy concerns related to attendance management.

8. Project analysis slide 3 SL No. Paper Title Authors Year Contributions Drawbacks 7 "A Comprehensive Study on Behavioral Biometrics for Cheating Detection in Examinations" Laura Anderson, Mark Wilson 2023 Investigates the application of behavioral biometrics, such as keystroke dynamics and mouse movement, for cheating detection in exams. Limited exploration of other biometric modalities, and may not address challenges specific to facial recognition for attendance. 8 "A Comprehensive Study on Behavioral Biometrics for Cheating Detection in Examinations" Laura Anderson, Mark Wilson 2023 Investigates the application of behavioral biometrics, such as keystroke dynamics and mouse movement, for cheating detection in exams. Limited exploration of other biometric modalities, and may not address challenges specific to facial recognition for attendance.

9. Project analysis slide 3 SL No. Paper Title Authors Year Contributions Drawbacks 9 "Machine Learning-based Analysis of Student Engagement for Cheating Detection" Rachel Bennett, Daniel Evans 2021 Investigates the use of machine learning to analyze patterns of student engagement as a means of identifying potential cheating instances during exams. May not provide a holistic view of cheating detection, and the effectiveness of engagement-based approaches may vary across different academic contexts. 10 "Biometric Data Security in Educational Environments" Samuel Clark, Patricia Adams 2020 Examines the security aspects of biometric data in educational settings, addressing concerns related to data storage, transmission, and protection. Primarily focuses on biometric data security, and may not provide detailed insights into the integration of biometrics with attendance and cheating detection.

10. Project analysis slide 3 EXISTING SYSTEM AND ITS DRAWBACKS Current examination monitoring heavily relies on manual invigilation and basic surveillance systems, which often fail to detect subtle or evolving forms of academic misconduct due to human limitations. Human invigilators may miss certain behaviors or may not be able to respond swiftly to emerging issues during examinations, leading to gaps in monitoring and potential instances of cheating going undetected. Conventional surveillance systems lack the capability to adapt to varying examination scenarios, making them less effective in preventing cheating and maintaining examination integrity. The shortcomings of the existing systems highlight the necessity for a technologically advanced solution that can autonomously and intelligently monitor examination halls. There is a necessity for a technologically advanced solution for autonomous monitoring

11. Project analysis slide 3 PROBLEM STATEMENT The increasing reliance on technology in examination processes has introduced new challenges related to maintaining the integrity of assessments. Instances of academic misconduct, such as cheating and unauthorized activities, pose a threat to the fairness and credibility of examinations. Traditional methods of invigilation are often reactive and struggle to keep pace with evolving cheating tactics. There is a pressing need for a proactive and intelligent system capable of detecting suspicious activities in real-time, ensuring a secure and reliable examination environment. Traditional invigilation methods tend to react to cheating incidents rather than proactively address them, making it challenging to keep up with the constantly changing tactics used by cheaters.

12. Project analysis slide 3 PROPOSED SOLUTIONS AND ITS ADVNATAGES The project employs a cutting-edge method to enhance examination integrity through the integration of machine learning and artificial intelligence (AI). Real-time video feeds from examination halls are analyzed using computer vision techniques, including facial recognition, gaze tracking, and object detection. A diverse dataset is utilized to train an ensemble of classifiers, incorporating deep neural networks for adaptability across various examination scenarios. Leveraging edge computing for on-site video processing ensures immediate responsiveness, with an integrated alert mechanism notifying invigilators of potential cheating. Continuous learning mechanisms refine the system's accuracy over time, offering a proactive solution to uphold examination integrity. PROPOSED SOLUTIONS

13. Project analysis slide 3 PROPOSED SOLUTIONS AND ITS ADVNATAGES Real-time Responsiveness: The system ensures immediate detection and alerting of suspicious activities, providing a swift response to maintain examination security. Continuous Learning: The incorporation of continuous learning mechanisms enhances the system's accuracy over time, allowing it to evolve and adapt to new tactics, thus staying ahead of potential misconduct during examinations. ADVANTAGES

14. Project analysis slide 3 ANALYSIS MODELLING (SYSTEM DESIGN) Student Attendance System Design Cheating detection System Design

15. Project analysis slide 3 SOFTWARE/HARDWARE TOOLS USED . 1. Software Requirements Operating system : Windows 7/8/10. Software Tool : Open CV Python Coding Language : Python Toolbox : Image processing toolbox Hardware requirements System : intel i3/i5 2.4 GHz. Hard Disk : 500 GB Ram : 4/8 GB Inbuilt or external camera

16. Project analysis slide 3 ARCHITECTURAL DESIGN System architecture using CNN

17. Project analysis slide 3 ARCHITECTURAL DESIGN State Chart Diagram

18. Project analysis slide 3 DETAILED DESIGN FLOW CHART OF OBJECT DETECTION FLOW CHART OF PREPROCESSING AND FEATURE EXTRACTION

19. Project analysis slide 3 DATA FLOW DIAGRAM . Data Flow Diagram of Smart Surveillance Solutions

20. Project analysis slide 3 DATA FLOW DIAGRAM . Data Flow Diagram for pre-processing module Data Flow Diagram for Object detection

21. Project analysis slide 3 USE-CASE DIAGRAM Use-Case Diagram .

22. Project analysis slide 3 SEQUENCE DIAGRAM . Sequence diagram of exam cheating detection

23. Project analysis slide 3 IMPLEMENTATION . METHODOLOGY

24. Project analysis slide 3 IMPLEMENTATION . Attendance Recognition

25. Project analysis slide 3 IMPLEMENTATION . Attendence Marking

26. Project analysis slide 3 IMPLEMENTATION . Cheating Detection Recognition

27. Project analysis slide 3 ALGORITHMS Different regions of face 1) Haar Cascade Algorithm Haar-Cascade Types Edge Features, Line Features, Four-rectangle Features Haar -cascade detection in OpenCV

28. Project analysis slide 3 ALGORITHMS 2) YOLO Algorithm The Convolutional neural networks have 4 layers • Convolutional layers • ReLU layer •Pooling layer •Fully connected layer.

29. Project analysis slide 3 ALGORITHMS

30. Project analysis slide 3 ALGORITHMS 3) Local binary pattern histogram (LBPH) Face Recognition Algorithm LBP =∑7 𝑠(𝑖𝑛 − 𝑖𝑐)2𝑛 Here, 𝑖𝑐 = 𝑐𝑒𝑛𝑡𝑒𝑟 𝑝𝑖𝑥𝑒𝑙𝑠 𝑣𝑎𝑙𝑢e 𝑖𝑛 =𝑁𝑒𝑖𝑔ℎ𝑏𝑜𝑢𝑟 𝑝𝑖𝑥𝑒𝑙𝑠 𝑣𝑎𝑙𝑢𝑒𝑠

31. Project analysis slide 3 ALGORITHMS

32. Project analysis slide 3 TESTING

33. Project analysis slide 3 TESTING

34. Project analysis slide 3 TESTING

35. Project analysis slide 3 TESTING

36. Project analysis slide 3 TESTING

37. Project analysis slide 3 PERFORMANCE EVALUATION Accuracy : Measure the accuracy of the AI model in detecting cheating instances correctly. This can be calculated by comparing the number of correctly identified cheating cases against the total number of cheating cases. False Positive Rate : Evaluate the false positive rate of the detection system, which indicates the proportion of non-cheating instances incorrectly flagged as cheating. A lower false positive rate is desirable to minimize unnecessary interventions. Speed : Assess the speed at which the AI system detects cheating instances. Faster detection allows for timely intervention and minimizes the impact of cheating. Robustness : Evaluate the robustness of the AI model against different cheating techniques and environmental conditions. A robust model should perform consistently well across various scenarios, such as different types of cheating behaviors or varying lighting conditions in exam halls.

38. Project analysis slide 3 APPLICATIONS Real-time Monitoring: AI can be employed to monitor exams in real-time, analyzing various data points such as eye movements, keyboard activity, or even facial expressions to detect signs of cheating as it happens. Plagiarism Detection: AI-powered systems can scan and compare exam answers with a vast database of existing content to identify instances of plagiarism or unauthorized copying. Anomaly Detection: AI algorithms can identify anomalies in exam responses, such as sudden spikes in correct answers or similarities among students' answers that suggest collaboration or cheating. Behavioral Analysis: By analyzing patterns of behavior during exams, AI can detect irregularities such as unusual pauses, repeated glances at unauthorized materials, or inconsistent response times, indicating potential cheating attempts.

39. Project analysis slide 3 CONCLUSIONS In this project, we hope to apply and investigate various human behaviour such as gazing out the window, conversing with people, focusing n other directions, moving about, and so on. We only utilize the CNN model because of its quicker object detection approaches available. In this project, it is the time to think for the exam to be digital as the world is growing at a very rapid pace and to compete with such, it is required to make examiner give their exam from their own location easily. Thus, this online proctoring system will help those examiners who actually live far away from from their college/university or to examination area where actually the exam needs to take place.

40. Project analysis slide 3 At the present approach, the teachers can only identify the malpractice and logs them off from the test. But it can be further developed into an application where certain period of malpractice can be recorded and evolution of the test can be done online and obtained marks can be sent to the student along with their malpractice while attending examination. It will be easy to evaluators by automatic updating results and malpractice. FUTURE ENHANCEMENTS

41. Project analysis slide 3 REFERENCES [1] . Zhang, Xiangyu , Xinyu Zhou, Mengxiao Lin, and Jian Sun. “ Shufflenet : An extremely efficient convolutional neural network for mobile devices.” In Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 6848-6856. 2023 [2] . Ahmad Salihu Ben-Musa, Sanjay Kumar Singh, Prateek Agrawal, “Suspicious Human Activity Recognition for Video Surveillance System”, International Conference on Control, Instrumentation, Communication and Computational Technologies, Research gate, 2022 [3] . B. C. Amanze , C. C. Ononiwu , B. C. Nwoke , I. A. Amaefule , “Video Surveillance And Monitoring System For Examination Malpractice In Tertiary Institutions”, International Journal Of Engineering And Computer Science, Vol. 5, January 2021 [4] . X. Wang, M. Xia, H. Cai , Y. Gao , C. Cattani , “Hidden MarkovModels -Based Dynamic Hand Gesture Recognition”, Mathematical Problems in Engineering, pp.1-10, 2020

42. Project analysis slide 3 REFERENCES [5] . R. Lockton , A.W. Fitzgibbon, “Real-Time Gesture Recognition Using Deterministic Boosting”, Proc. British Machine Vision Conference, pp. 817-826, Sept. 2020 [6] . Paul Viola, Michael J. Jones , “Robust Real time Face Detection, International Journal of Computer Vision”, Vol 57, issue 2, pages: 137-154, 2019 [7] . Richard Szeliski , “Computer Vision: Algorithms and Applications”, First Edition, Springer, 2019 [8] . Teddy Co, “A Survey on Behavior Analysis in Video Surveillance for Homeland Security Applications”, IEEE Conference on Applied Imagery Pattern Recognition Workshop (AIPR), Pages: 1 – 8, 2019

43. Thank You