MONITORING EMPLOYEES ENTERING AND LEAVING THE OFFICE WITH DEEP LEARNING ALGORITHMS

Abstract

The importance of face recognition systems has sped up in the last few decades. A face recognition system is one of the biometric information processing, but its applicability is easier and the working range is larger than other processing, i.e.; fingerprint, iris scanning, signature, etc. This study attempts to create a system to monitor employees entering and leaving the office using face recognition. In addition, the system also signals by LED when recognizing a staff who has clearance to enter or notifies those who do not in the area. Events of entering and leaving from staff are written into a log file for management purposes. The face-detection and image preprocessing utilize Multi-task Cascaded Convolutional Network. Feature data is then extracted from the processed images by FaceNet, which is classified by the Support Vector Machine algorithm into a model. Information of employees and logs are saved in MySQL database, which is also used in a web application using Python and Django web framework. Two doors at Cantho University Software Center office were selected for the research. Data were collected from 4 people with their information and face photos. The collected information was saved into the system’s database while its face recognition model learned the face photos. Besides this, further employees can also easily register to be recognized by the system. The tested system has acceptable performance to recognize faces and is also capable of detecting and recognizing multiple faces in live acquired images.