Optimized Student's Multi-Face Recognition and Identification Using Deep Learning
Avadhoot Autade, Pratik Adhav, Abhimanyu Babar Patil, Aditya Dhumal, Sushma Rahul Vispute, K. Raja Rajeswari · 2023
Numerous real-life applications that require human identification inspire face recognition research. Many authors have employed deep-learning-based face recognition applications for security purposes, such as attendance tracking systems and person verification. Person identification with deep learning for threat avoidance and criminal detection succeeded well, according to the survey. Numerous authors claim that this system’s accuracy ranges from 96 to 99%. According to onesurvey, this deep learning model performs better than the other models with an accuracy of 95–98% for the attendance monitoring system. A common finding of the survey shows that research is done for three categories of work:first, person identification for security purposes; second, student face recognition from a single face image dataset; and third, student face recognition from a multi-face image dataset. In this research, the authors propose an optimized tolerance based multiple face recognition model to display the results for student face recognition. This model’s accuracy rate is 99.38% for single face image data and 98% for multiple face image data with a tolerance value 0.5.