The system of monitoring students’ behavior during e-tests

Abstract

Problem. In connection with the transfer of the educational process to a fundamentally new level, the role of independent work of students, which has become the main form of the knowledge, is growing significantly. Therefore, the computer testing is becoming increasingly important as a means of monitoring the students' learning materials. These speedsup the control process and ensures its objectivity. However, the testing process is vulnerable to possible students’ abuse – use of aids, outside help, etc. Goal. The goal is developing a client-server system for analyzing students' behavior during e-testing. Methodology. Solution of the set tasks involved application of the set-theoretical approach in order to build up the model of an observation process for students' behavior and their component organization as well as their interaction, image processing methods. Results. A model of the monitoring process of the students' behavior is proposed, which allows real-time monitoring of violations during e-testing based on face recognition. The principle of operation of the developed system also described comparative analysis with existing analogues. Originality. A monitoring system is proposed, which includes the following features: tracking the active URLs, obtaining the list of network connections, real-time tracking of any changes in parameters (width and height) of the active window, creation of screenshots of the working area of the computer screen and photographs using a web camera, automatic reading of the contents of the clipboard and keeping a log of pressed keys, taking screenshots of the working area of the computer screen and photographs using a web camera, recognizing the student's face and tracking their position regarding the webcam. If violations are detected, the received messages are recorded and sent to the server. Practical value. The use of the proposed monitoring system for students' behavior will allow to remotely identify a larger number of violations that were not previously detected.