ARTIFICIAL INTELLIGENCE-BASED SMART CLASS ATTENDANCE SYSTEM: AN IOT INFRASTRUCTURE

Attending students in many universities' lectures is still done following the traditional way, which is by passing an attendance sheet to be signed by the students or calling the students' names. Lecture time can be decreased considerably by following conventional attendance methods. This research aims to design, build, and demonstrate an IoT system of systems that automatically detects students attending a class within a classroom and updates an attendance database. In the proposed system, the system of interest is designed to detect each student entering the classroom at the doorway using a camera. This ensures, in almost all cases, the student entering the classroom is facing the camera. Then, the system of interest identifies students using an artificial intelligent method by utilizing a facial recognition technique. This way tilled or side-face images is reduced for better and faster recognition. After, the system of interest updates the roll based upon the recognition process. Nevertheless, in the case of unrecognition of any participant, the proposed system will display the total sum of the unrecognized people on a screen along with an illustration of the attendance sheet. Finally, the proposed system was tested and evaluated several times; and it has been proven it is trustable enough since its recognition accuracy is around %93 on average.