<![CDATA[Employee attendance is a crucial aspect of human resource management, particularly in maintaining discipline and ensuring the operational effectiveness of a company. PT KAMM currently uses a fingerprint-based attendance system which, although effective, often encounters issues such as sensor sensitivity to finger conditions, potential device damage caused by continuous physical contact, and employee inconvenience. This research aims to develop a face recognition-based attendance system as a more efficient and hygienic alternative. The dataset comprises 1,400 facial images from 20 PT KAMM employees (20 classes), split into 80% training, 10% validation, and 10% testing data. The method applied combines the Haar Cascade algorithm for face detection and a Convolutional Neural Network (CNN) for face recognition. The CNN architecture consists of four convolutional layers with 32 to 256 filters, ReLU activation, max pooling, flatten, a 512-neuron fully connected layer, dropout of 0.5, and softmax classification. The model was trained for 50 epochs using the Adam optimizer with a learning rate of 0.001 and batch size of 32. Evaluation was conducted using accuracy, precision, recall, and F1-score metrics. Results show the system achieved an accuracy of 95.71%, precision of 95.80%, recall of 95.60%, and an F1-score of 95.70%, with an average inference time of 0.12 seconds/frame in real-time. However, the system has limitations: accuracy drops by up to 12% under extreme lighting conditions and when employees wear masks. This study is expected to serve as a reference for other companies seeking to adopt similar face recognition technology for contactless attendance management systems.]]>
Copyrights © 2026
