Visual Analytics for Online Exam Proctoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current online proctoring methods, particularly fully automated approaches, face challenges due to the 'black box' nature of machine learning algorithms and biased training datasets, while semi-automated methods lack a convenient way for proctors to analyze suspected cheating behaviors and require costly, multiple devices for students during online exams.
Innovation Solution
A visual analytics system that combines client-side computing and video interaction data analysis to detect abnormal behavior, providing multi-level visualizations for reviewers to assess cheating, reducing the need for multiple devices and enhancing the efficiency of proctoring by using camera and mouse movement data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If fully automated proctoring approaches using machine learning techniques are used, then the usage of manpower is eliminated, but the system suffers from black box nature and unreliable decision-making due to biased training datasets
Solution Approach 1:
The patent introduces visual analytics as an intermediary layer between automated detection and final proctoring decisions. This intermediary provides transparent, interpretable visual representations of detected behaviors, allowing reviewers to understand the basis of automated detections while maintaining human oversight for final decisions.
Solution Approach 2:
The patent segments the proctoring system into multiple components: automated detection module, visual analytics module, and human reviewer module. This segmentation allows each component to perform its specialized function while maintaining overall system reliability through distributed decision-making.
2Reliability
If semi-automated proctoring approaches are used to involve humans in final decision-making, then decision reliability improves, but proctors lack convenient ways to explore and analyze suspected cheating behaviors
Solution Approach 1:
Visual analytics serves as an intermediary tool that bridges automated detection results and human reviewer understanding. It provides intuitive visual representations that make complex behavior data easily explorable and analyzable by proctors.
Solution Approach 2:
The patent replaces manual exploration of raw data with automated visual analytics processing. Instead of proctors manually examining raw video or sensor data, the system automatically generates visual representations that highlight suspicious behaviors and patterns.
3Measurement precision
If existing semi-automated proctoring technologies require multiple devices (two webcams, gaze tracker, EEG sensor) to record exam process, then detection capability improves, but the cost becomes unaffordable for most educational institutions
Solution Approach 1:
The patent makes the single webcam multi-functional by combining computer vision algorithms with mouse movement analysis. Instead of requiring separate devices for different detection functions, the system uses one device to perform multiple detection tasks through software processing.
Solution Approach 2:
The patent replaces physical sensor devices (gaze trackers, EEG sensors) with software-based alternatives using computer vision and interaction data analysis. This substitution maintains detection capability while eliminating the need for expensive specialized hardware.
Data Source
AI summary
A system for proctoring online exams includes a client-side computing system and a visual analytics system. The client-side computing system includes a camera configured to obtain video data corresponding to a user while taking an exam and one or more input devices configured to obtain interaction data, wherein the interaction data includes mouse movements of the user while taking the exam. The visual analytics system is configured to obtain the video data and the interaction data from the client-side computing system, analyze the exam data to detect abnormal behavior by the user based at least in part on mouse movement data, and generate one or more visualizations of the analyzed exam data to be used in determining whether or not the user has cheated during the exam.