Automated Attendance System Using Facial Recognition

Abstract

Automatic Attendance System has the potential to simplify attendance management procedures in many educational contexts, the development of an automatic attendance system utilizing facial recognition has attracted considerable attention in recent years. The system described in this paper uses a camera installed in a classroom to take pictures of students as they participate in class. The Multi-Task Cascaded Convolutional Networks (MTCNN) technique is used to analyze the collected images in order to identify and extract facial features. A model built using transfer learning methods is used to enable facial recognition. With the aid of transfer learning, the model can improve its capacity for face recognition and classification by making use of prior knowledge from a sizable dataset. The model is learned using a deep learning framework [10], allowing it to pick up on complex facial patterns and features. The system runs in a loop throughout the class period, utilizing the camera to continuously take pictures of students' faces. In the recorded photos, facial regions of interest are found and aligned using the MTCNN algorithm. The trained facial recognition algorithm uses these aligned faces as input and compares the derived attributes to an existing database of registered pupils to identify the subjects. The model gives identities to the observed faces, making it possible to track attendance accurately and effectively. The suggested system has a number of benefits, such as real-time attendance tracking, less administrative workload, and improved attendance management accuracy. Automating the attendance process allows educators and institutions to devote more time to teaching activities, improving the learning process as a whole. The suggested system's experimental assessments show encouraging results, with high accuracy rates in face detection and identification tasks. The system's effectiveness could be impacted by issues such varying lighting conditions, posture, and occlusion. To address these issues and raise the robustness and reliability of the system, more investigation and improvement are needed.