Enhanced Criminal Identification Using Integrated Face Recognition and Blockchain Technology
R Sreelakshmi, Shanigacharla Ajay, P. Pavani, Bidari Vamshi, Bura Saleem · Journal of Sensors, IoT & Health Sciences (JSIHS). · 2025
This paper presents a novel Blockchain-Based Face Detection and Recognition System aimed at enhancing criminal identification through the integration of deep learning algorithms and blockchain technology. The system leverages an Integrated Deep Learning (IDL) model for high-accuracy facial recognition under varying conditions such as occlusion, lighting variations, and background noise. To ensure data integrity, transparency, and immutability, every recognition result is securely recorded on a blockchain ledger, preventing unauthorized access and tampering. The proposed framework addresses key challenges in conventional surveillance systems, such as centralized vulnerabilities and data manipulation, by providing a decentralized, tamper-resistant architecture. Experimental results demonstrate high performance in terms of accuracy, precision, recall, and F1-score, even in challenging environments The proposed model was tested under diverse real-world conditions, including frontal faces, occlusion, side angles, lighting variations, and spoofing attempts. The system achieved a maximum recognition accuracy of 99.1%, with precision reaching 98.7%, recall at 99.3%, and an F1-score of 99.0% during real criminal identification scenarios. Even in challenging environments, such as low lighting and background noise, the model maintained accuracy above 89%, demonstrating robustness. Spoof detection was successfully handled with 98.3% accuracy and only 1.4% false positive rate, showcasing the system's resilience against fraudulent attempts. Each recognition event is securely logged on a blockchain ledger, ensuring data immutability and traceability, with an average logging time of under 480 milliseconds.