Indian Journal of Science and Technology

Abstract

Background/Objectives: The objective of this paper is to introduce an automated attendance marking system for LPU using CCTV cameras. It detects the students sitting in the class and takes the attendance automatically without any human assistance and count total number of students present in the class also. Methods/Statistical Analysis: An enhanced detection algorithm is proposed that is having two phases: color image processing of RGB planes and modified background subtraction of image which executed simultaneously for object detection. Before detection process, noise is removed from images by using log function which uniformly distributes the brightness over the image. To count the number of detected objects, a counter is used which calculates total number of objects present in an image. The Automatic ROI detection algorithm is also proposed which detects region of interest automatically on the basis of repeated regular pattern of object. Findings: The system able to count and mark attendance of students sitting in class room. Implementation shows that our proposed approach has 89.5% accuracy with less false positive and false negative rate as compared to other traditional approaches. The Automatic ROI detection algorithm also detects the region of interest with considerably high accuracy. Application/Improvements: The proposed system will easy to deploy in various schools, colleges, universities and other educational institutes with minimal setup cost where CCTV cameras are already installed in class rooms. 
Keywords: Background Subtraction, Log Function, Object Detection, RGB Color Image Processing, Region of Interest