Security of facial biometric authentication for attendance system

Face image processing has become one of the fields of computer vision in processing computerized image patterns; the face becomes one of the vital biometrics that stores essential information used in predicting the characteristics of a person. Biometric techniques with facial recognition systems are now required in various fields, including business, one of which is the attendance marking system that is a crucial repetitive transaction requirement because it relates to employee productivity. In terms of ethics, attendance recording using a person's face has many benefits; one of them is removing the necessity to make direct contact with the scanning device. Before doing face recognition, one of the preprocessing stages is face detection as an effort to find the existence of a face image consisting of eyes, nose, mouth, and other facial features. This research employed Viola-Jones method for face detection, Gabor Wavelet for feature extraction, and Template Matching. Two scenarios are applied for attendance recording, individual face recording, and group face recording where several faces are captured simultaneously, and each face is extracted and recognized. For Individual attendance recognition, this research achieved an accuracy of 75%, recall 64%, and precision of 88%. The better result is shown on simultaneous/group face recognition, and the research achieved 88% accuracy, 75% of recall, and 97% of the precision score.

Automated summary

This study focuses on using biometric techniques and facial recognition systems for attendance marking, achieving high accuracy in both individual and group face recognition. Good results were obtained using Viola-Jones method for face detection, Gabor Wavelet for feature extraction, and Template Matching in face recognition.