Deep Learning Approach for Facial Identification for Online Transactions
Abhinav Gill, Dhyanendra Jain, Jatin Sharma, Ashish Kumar, Puneet Garg
Abstract
The “Deep learning approach for facial identification for online transactions” project represents a pioneering effort in the domain of secure and efficient user verification. This project innovatively integrates Aadhar data, a comprehensive repository of biometric and demographic information, with advanced facial recognition technology. Employing the Local Binary Pattern Histogram (LBPH) algorithm, the system achieves real-time face recognition, ensuring a seamless and rapid authentication process. The amalgamation of Aadhar data significantly enhances the accuracy and reliability of user identification. Stringent security measures, including robust encryption techniques, are implemented to safeguard sensitive biometric information, addressing privacy concerns and aligning with stringent data protection laws. Pilot testing of the system underscores its user-friendly interface, prioritizing accessibility without compromising security. The project delves into the comparative analysis of face recognition algorithms, spotlighting the efficacy of LBPH in real-time scenarios. Discussions encompass the legal and ethical implications of utilizing Aadhar data, emphasizing compliance with evolving data protection regulations and ethical standards. Challenges encountered during the project, such as potential biases in facial recognition algorithms, are acknowledged, paving the way for future research directions focused on bias mitigation and algorithmic refinement. The scalability of the system is demonstrated through success.