Motion-Activated Facial Recognition-Based Electronic Residential Logbook using PCA Algorithm with Web Application

摘要

This paper embarks on the development of an electronic residential logbook that is motion-activated and utilizes facial recognition as its biometrics authentication. This paper aims to access log information through a web application. The algorithm that was used for the facial recognition process is known as the Principal Component Analysis, a linear technique that is widely used in recognition and compression applications. Moreover, an application of PCA which is the Eigenface technique was used in capturing the most necessary factor in the facial recognition process. The hardware part of the electronic residential logbook is driven by AT89S52 microcontroller where a PIR sensor is connected. The PIR sensor is responsible for the motion activation of this design. The log record information can be viewed through a web application which reflects the time in/out of the residents. In conclusion, the researchers were able to meet all of their objectives and provided the expected output with a 13% error rate for the facial recognition system with only 5 training faces conducted for each resident.