DEEP LEARNING-DRIVEN FACE RECOGNITION ATTENDANCE APP
| dc.contributor.author | JUSTINE LING KE YI | |
| dc.date.accessioned | 2026-04-20T06:53:18Z | |
| dc.date.issued | 2025 | |
| dc.description | Traditional attendance methods, such as manual signing in or swiping cards to punch in, are usually inefficient, error-prone, and risk cheating. To solve these problems, this project proposes an artificial intelligence-based face recognition attendance system and deploys it in an Android mobile application. The system employs deep learning techniques, in particular a metric learning architecture based on Convolutional Neural Networks (CNN), for efficient extraction and comparison of face features. This study compares several classification methods, including a combination of PCA with Random Forest, SVM, KNN and LDA, and the results show that the CNN model achieves the highest recognition accuracy of 88.51% on the LFW dataset. After training, the model was converted to TensorFlow Lite format and successfully integrated into an Android application, achieving real-time recognition and a convenient attendance recording interface. Software tests verified the stability and usability of the system. Although the current system is already usable, the future plan is to integrate the map API to support the geolocation-based verification function to further enhance the user experience. This project validates the feasibility of deep learning-based face recognition in attendance management and provides a secure, efficient and scalable solution for educational institutions. | |
| dc.identifier.uri | https://scholarhub.unimas.my/handle/123456789/381 | |
| dc.language.iso | English | |
| dc.publisher | Universiti Malaysia Sarawak (UNIMAS) | |
| dc.relation.ispartofseries | Faculty of Computer Science and Information Technology | |
| dc.subject | driven face recognition | |
| dc.title | DEEP LEARNING-DRIVEN FACE RECOGNITION ATTENDANCE APP | |
| dc.type | Final Year Project |