AI-Based Criminal Fraud Detection System Using Facial Recognition

This paper presented an AI-Based Criminal Fraud Detection System that integrates deep learning facial recognition, criminal database management, real-time alert generation, and a law-enforcement-oriented Streamlit interface into a single, locally deployable platform. The system demonstrates that automated facial recognition can meaningfully improve upon manual surveillance in accuracy, speed, consistency, and scalability. The 85.7% overall recognition accuracy, 5.1% false-positive rate, and sub-2-second identification speed validate the core thesis that AI-based systems can serve as effective decision-support tools for law enforcement. The lower false-positive rate relative to human operators is particularly important for protecting the civil liberties of innocent individuals in high-stakes identification contexts. The system's most significant practical limitation is environmental sensitivity, particularly to lighting conditions and extreme facial angles. Future work will address these through infrared camera integration, multi-view enrollment, and advanced preprocessing pipelines. Planned enhancements also include cloud migration for multi-site deployment, approximate nearest-neighbor search for large-scale database performance, mobile companion applications for field personnel, and multi-modal biometrics incorporating gait and voice recognition. The project validates the broader thesis that AI-based facial recognition, when implemented with appropriate user interfaces, database management, and accountability mechanisms, can be developed into a practical and ethically responsible tool for modern criminal fraud detection and surveillance applications.

[1] Kumar A., Mehta R., Singh P. (2023). Real-Time Face Recognition Using Deep Learning and OpenCV. IEEE Transactions on Image Processing, Vol. 32, pp. 1847–1862. DOI: 10.1109/TIP.2023.3245678 [2] Davis J., Robert L. (2022). Face Recognition using HOG and Deep Learning-based Embeddings. International Journal of Computer Vision (IJCV), Vol. 130, Issue 8, pp. 2134–2156. DOI: 10.1007/s11263-022-01634-4 [3] Sharma A., Gupta V. (2024). Criminal Identification using Automated Facial Recognition Systems. Expert Systems with Applications (Elsevier), Vol. 238, Article 121893. DOI: 10.1016/j.eswa.2023.121893 [4] Patel R., Sharma K. (2023). Automated Surveillance Using Computer Vision. Journal of Computer Vision, Vol. 45, Issue 3, pp. 789–812. [5] Mehta S., Kumar A. (2024). Privacy-Preserving Facial Recognition for Law Enforcement. IEEE Security & Privacy, Vol. 22, Issue 1, pp. 34–47. [6] Bradski, G. (2000). The OpenCV Library. Dr. Dobb's Journal of Software Tools. Available at: https://opencv.org [7] Taigman, Y., Yang, M., Ranzato, M., Wolf, L. (2014). DeepFace: Closing the Gap to Human-Level Performance in Face Verification. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1701–1708. [8] Schroff, F., Kalenichenko, D., Philbin, J. (2015). FaceNet: A Unified Embedding for Face Recognition and Clustering. CVPR 2015, pp. 815–823. [9] He, K., Zhang, X., Ren, S., Sun, J. (2016). Deep Residual Learning for Image Recognition. CVPR 2016, pp. 770–778.