Smart Attendance System Using Deep Learning

EasyChair Preprint 7441

Abstract

Engagement is the key to the success of the intelligent user interface. Such interface requires to respond appropriately and also needs to recognize the level of engagement. This paper presents a Deep Learning model to improve face recognition by implementing and comparing various methods used for robust face detections where the subject would give attention to the interface for a few seconds or the subject does not have the attention of the interface or just passes by the interface to register an engagement. By combining some powerful deep learning tools like Convolution Neural Network with techniques like Histogram of oriented Gradients, we can out form the older techniques and with more accuracy and with an computational efficiency ,the model is trained by CNN in order to precisely measure the embedding of an face (128 measurement of each face),the network was trained by Deep Learning and using any simple classifier and find the closest match with the measurements received from the model with the images present in the database. The end result will be the name of the student being recorded for the respective use.

Keyphrases: CNN, HOG, deep learning, face recognition

@booklet{EasyChair:7441,
  author    = {Atharva Amrapurkar and Pratik Parbat and Vandana Jagtap},
  title     = {Smart Attendance System Using Deep Learning},
  howpublished = {EasyChair Preprint 7441},
  year      = {EasyChair, 2022}}