Research On The Key Technology Of Candidates' Attendance In Intelligent Examination Room

Intelligent behavior analysis is a hot issue in the field of computer vision, because of the uncertainty of the target behavior and the complexity of the external environment, it is hard to put forward a general method of video object analysis and behavior understanding, but it can be analyzed and understood according to the specific situations and the characteristic of video objects.This paper is aimed at the study of the students’ attendance which is a specific application of intelligent behavior analysis in the intelligent examination room, because examinees keep a relatively stable examination room environment in the exam, so can recognize and analysis whether the examinee is in the examinee’s place and has leaving behavior. To implement the intelligent of the students’ attendance management can strengthen national education examination management and to maintain the fairness and justice, security, and examination order has an important significance. In this paper, the main research work as follows:(1) The intelligent design of video surveillance system. As the limitations of video surveillance system, this paper analysis and design the totality framework to the intelligent of video surveillance system based on the purpose of the construction of standardized examination room.(2) The matching between physical examination and simulation examination. First of all, original image after the perspective transformation can generate orthograph, then using the template matching to get the desk area, finally calculate Euclidean distance between the clicked desk area of supervisor in physical examination room and the center of each desk area to identify the rows and columns which desk belonging to, so can achieve the matching between physical examination and simulation examination.(3) Candidate’s recognition based on similarity matching. This paper firstly get examinee’s area according to the desk’s area, and then realize the recognition of examinee’s reign through detecting whether has someone in the examinee’s reign by using the normalized correlation matching and background subtraction. The experiments prove that combine these two methods can improve the accuracy obviously. This paper solves the problem of the examinee block before and after by using the order matching and filling the background.(4) Student’s leave recognition based on moving target tracking. This paper proposed the particle filter algorithm based on HSV color histogram for tracking the examinee to achieve the purpose of students leave identification, experiments prove that this method has higher accuracy in identifying candidates leave which is abnormal behavior.(5) The comprehensive experiment of the key techniques of candidates in attendance in the intelligent examination. Through the experiment algorithm parameters or threshold was chosen, this paper show and analysis the algorithm results.