<![CDATA[With the advancement of increasingly affordable camera technology and the ease of image capture processes, facial recognition has become the biometric authentication method most vulnerable to spoofing attacks. This vulnerability significantly undermines the data integrity and reliability of facial recognition-based attendance systems. Therefore, a liveness detection method is required to help mitigate these spoofing attacks. This study aims to enhance the reliability of existing attendance systems in the teacher and staff attendance application developed for Universitas Mikroskil. In this research, the reliability of the liveness detection system will be tested using a nonlinear diffusion algorithm and Convolutional Neural Network (CNN) against spoofing attacks, employing a confusion matrix method. Based on the conducted tests, the designed liveness detection system achieved an accuracy rate of 87.76%.]]>
Copyrights © 2025
