MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202611058187 A) filed by Ashish Kumar Mathur; Atharv Kishor; Abhyudaya Rajbhar; and Aditya Sharma on May 07, 2026, for Ai-Based Proctoring System For Online Tests.
Inventors include Ashish Kumar Mathur; Atharv Kishor; Abhyudaya Rajbhar; and Aditya Sharma.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The rapid expansion of online education and digital assessments worldwide has created an urgent need for robust, scalable, and intelligent examination integrity solutions. This patent discloses an Artificial Intelligence-based Online Examination Proctoring System that employs a multi-modal surveillance architecture to automatically detect, record, and report suspicious behaviour during remote examinations — without requiring continuous human involvement. The disclosed system integrates computer vision, deep learning, and audio analytics to provide real-time automated invigilation. It performs face recognition using the Local Binary Pattern Histogram (LBPH) algorithm and MTCNN/FaceNet models for identity verification both at the commencement and throughout the duration of the examination. Object detection using the YOLOv3/YOLOv5 deep neural network model identifies prohibited items such as mobile phones, unauthorized persons, and foreign objects. Gaze and head-pose estimation via MediaPipe and OpenCV tracks thedirection of a student's attention, flagging sustained off-screen gazes. Audio analysisusing WebRTC Voice Activity Detection (VAD) detects abnormal sounds, background conversations, or multiple voices indicative of verbal cheating. Browser tab-switch monitoring detects attempts to access unauthorized digital resources. The system is built on a Django-based backend with REST/WebSocket APIs andaPostgreSQL database, providing secure, time-stamped malpractice logging, real-timealert generation, post- exam integrity reporting, and an administrator dashboard. Afault-tolerant session recovery mechanism allows examinations to be resumedfromthe last saved state after power outages or connectivity failures. The systemsupportscross-platform deployment on laptops, desktops, and mobile devices. Keywords: Artificial Intelligence, Online Proctoring, Face Recognition, LBPH, YOLOv3, Gaze Tracking, MediaPipe, Audio Analytics, Academic Integrity, DeepLearning, Computer Vision, Django, WebRTC VAD, Remote Examination.
Disclaimer: Curated by HT Syndication.
