Badri, Syahrul (2024) Sistem Absensi Dengan Pengenalan Wajah Menggunakan Metode Histogram Of Oriented Gradients (hog) Dan Support Vector Machines (svm) Berbasis Website. Other thesis, Universitas Islam Riau.

Text 203510320.pdf - Submitted Version Restricted to Registered users only Download (9MB) | Request a copy

Abstract

Student attendance in class is one of the important indicators of academic activity and achievement. However, the manual attendance method has various weaknesses, such as the potential for cheating and errors in data recapitulation. This study aims to develop a website-based attendance system with facial recognition using the Histogram of Oriented Gradients (HOG) and Support Vector Machine (SVM) methods. The test results show that at a distance of 30 cm, the False Acceptance Rate (FAR) and False Rejection Rate (FRR) levels reach 0. At a distance of 50 cm, the FAR value increases to 0.1 (10%), while the FRR remains 0. The optimal distance for facial recognition is between 50 cm and 100 cm. At a distance of 100 cm, the FAR value increases to 0.3 (30%), while the FRR increases to 0.2 (20%). This study concludes that the attendance system with HOG and SVM-based facial recognition has optimal performance at a certain distance. For further development, this system is expected to be integrated with Internet of Things (IoT) based technology and use deep learning methods to improve accuracy and flexibility

Actions (login required)

View Item