Online Examination Auto-Proctoring System

One of the biggest challenges with online examinations today is the identity verification and proctoring of students giving online exams. For examination, the authority must confirm the authenticity of the online students who are attempting the exam. Additionally, they must make sure that the students successfully complete all of the online examinations without cheating or using any inappropriate methods. So here we have proposed a detailed explanation of the Auto-Proctoring methodologies used in online examination. Currently, human proctoring is the most common assessment method, requiring teachers and students to join online meetings, where they are visually monitored via webcam. However, these methods are laborious and expensive. This system includes a face recognition and attendance system to estimate user authentication developed using LBPH and SVM classifier. Further a talking detection algorithm based on mouth opening algorithm is implemented using Viola Jones algorithm and Mouth Aspect Ratio calculated using Facial Landmarks library by dlib. Audio detection method comprises Speech to text conversion using Speech Recognition libraries and API along with Text summarization using NLP. Further to prevent cheating, tab switching detection and Two or more person detection or No face detection is implemented in the system. By integrating these advanced methodologies and technologies, the proposed AutoProctoring system establishes a secure and cheating-free examination environment.