Vision-based ID wearing detection system using IP camera

• Main Authors: Paramio, Crischelle Mae R., San Jose, Jen Louie R.
• Format: text
• Language: English
• Published: Animo Repository 2016
• Subjects: Image processing—Digital techniques; Identification—Equipment and supplies; Electrical and Electronics; Systems and Communications
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/5493
• Institution: De La Salle University

For security purposes, many institutions adapt an ID wearing policy. In the case of academic institutions, the number of discipline officers is less than the number of students. With this, instances of not wearing an ID card might be inevitable. This study aims to develop a vision-based ID wearing detection system using IP cameras. This is one of the methods to check and monitor students whether or not they are wearing their IDs while they are in the campus. This study is significant in aiding the Discipline Officers in the case of DLSU-STC in monitoring students at different locations like the Ground Floor East Entrance, the Ground Floor East Canopy and the 2nd Floor Lobby of Milagros R. Del Rosario Building. The system uses one internet protocol (IP) camera implement a motion detection algorithm for detecting the moving objects, Linear Binary Patterns (LBP) for face detection and object (ID card) detection, and object annotation procedures to gather training data sets. The researchers considered two approaches: Contour using Motion Detection and Contour using Region of Interest.Results show that the first approach, the Contour using Motion Detection algorithm, has a better result accuracy compared to the second approach, the Contour using Region of Interest, since the first approach looks for the ID card if there is a detected motion while the second approach looked for the ID when the contour is formed. Accuracies were calculated per location resulting to a range between 93% to 96%.