Deep Residual Learning based Attendance Monitoring System
Ashok Kumar Tammisetti, Keerthi Sri Nalamalapu, Srija Nagella, Khwaza Shaik, Khaleel Ahmed Shaik · 2022 8th International Conference on Advanced Computing and Communication Systems (ICACCS) · 2022
In the field of personal identification Face recognition technology is more trustworthy than other technologies. In the proposed work, we elaborated on the use of face recognition in the monitoring of attendance of a class without the intervention of humans. Face recognition gives fruitful results compared to previous existing techniques like a fingerprint, iris, and RFID techniques. Using the capabilities of the dlib framework, this project provides a method for detecting and identifying the faces of students present in a classroom in real-time. Our trained network uses the CNN or HOG method to identify and locate students' faces in the captured frames. The facial features are extracted from those images using Deep Residual Networks called ResNets. And distinguishes between feature vectors using the KNN classifier. This network has an accuracy of 99.83% on the Labelled Faces in the Wild benchmark. And finally, we have shown some performance metrics and plots by comparing them with other existing methods in face recognition. Hence this method is cost-efficient and requires little maintenance.