Face Recognition-Based Attendance System Using Haar Cascade: A Case Study Between Siblings

Traditional attendance systems have long been recognised as inefficient and outdated. For the past years, a lot of research and implementation have been using technological attendance such as web attendance, QR code attendance and mobile attendance. However, due to technological limitations, complexity, inflated costs, power consumption, and limited data transfer capabilities, face recognition systems were not viable options. Today, with significant advancements in technology, computers have become more efficient and powerful, marking a notable departure from the limitations of the past. In UiTM, currently, the implemented system is either web attendance or the traditional attendance system. Many students have pointed out the issue with the web system as students with bad internet reception or issues with their mobile devices will have an issue recording their attendance. On the other hand, traditional systems have imposture issues, by having other people sign the attendance. In this paper, face recognition technology in attendance systems is explored with the benefits it offers, such as improved efficiency and accuracy. The method of the study is by using Python programming language in PyCharm with the implementation of Haar Cascade Adaboost Frontal Face Detector,-Based on the test results, the accuracy for individual testing is calculated at 59.72% with 26 total true positives and 17 true negatives, while the precision, positive rate and F1-score were 57.77
^%
, 0.7222 and 0.6419 respectively. In the group testing, the subjects between three siblings, on the other hand, accuracy, and precision both went down to only a mere 50%.