Classifying student conditions based on facial expressions is one application of Artificial Intelligence (AI) and Computer Vision that can be used as a tool to assist in observing learning activities. This study aims to develop a prototype system for real-time student condition classification using a Convolutional Neural Network (CNN) with a Transfer Learning approach. The system is designed to detect faces and classify three categories of student conditions: serious, chatting, and sleepy, based on video from a webcam. The research process includes collecting a video dataset, extracting frames, labeling data, image preprocessing, face detection using YOLO, and classification using a CNN based on Transfer Learning. The dataset is divided into training, validation, and testing data. Model evaluation is done using Top-1 Accuracy and inference time measurement. The test results showed that the model achieved a Top-1 Accuracy of 100% with an average inference time of 7.4 seconds per frame. The prototype can perform face detection and student condition classification in real-time through a webcam. The categories used in this study are operational categories based on learning conditions and are not intended as psychological diagnoses.
Copyrights © 2026