APPLICATION OF MULTI-TASK CASCADED CONVOLUTIONAL NEURAL NETWORK ALGORITHM IN SCHOOL SUPERVISOR ATTENDANCE SYSTEMS IN THE FIELD OF COMPUTER VISION

The attendance system used by the Education Department can be said to be still manual. Where they use the Timestamp application to take photos. Where the application only takes faces without detecting the face. Therefore, researchers created a face detection presence system by applying the Multi-Task Cascaded Convolutional Neural Network algorithm using the face-api.min.js library for the face detection process. The aim of this research is to make it easier for school supervisors to manage attendance, so they can provide accurate information. Then, based on the research results, a face detection and location detection system for school supervisors was successfully developed using the Multi-Task Cascaded Convolutional Neural Network (MTCNN) algorithm. From the results of tests carried out using a dataset of 140 images from 28 people with different photos taken (face view, top view, bottom view, left side view, right side view). Test results on the facial presence detection system using the MTCNN (Multi-Task Cascaded Convolutional Neural Network) algorithm succeeded in detecting faces by 100%.