Face Recognition Based Attendance System

EasyChair Preprint 10070

Abstract

This paper discusses an attendance system machine learning techniques like CNN, LBP and HoG is proposed. The system comprises an equipment package that includes a camera constructed to capture images along with an independent server set up specifically tailored towards data processing activities alongside storage functions respectively. In this way incoming photos taken are then analyzed and thereby matched against pre-existing profiles in order to produce precise identification outputs immediately without manual registration complexities nor dependence on physical ID card usage – subsequently transforming traditional methods into streamlined procedures across multiple levels effectually and efficiently optimizing resources further ensuring proper accountability entirely within organizations currently under test conditions proposed. The system when tested with classifiers attained an accuracy of 98.4% over the existing datasets, thus exceeding prior benchmarks.

Keyphrases: Back Propagation, Overfitting, convolution, face detection, face recognition

@booklet{EasyChair:10070,
  author    = {Saachi Shrikhande and Siddhesh Borse and Shripad Bhatlawande},
  title     = {Face Recognition Based Attendance System},
  howpublished = {EasyChair Preprint 10070},
  year      = {EasyChair, 2023}}