MUMBAI, India, Feb. 13 -- Intellectual Property India has published a patent application (202621000212 A) filed by Dr. Neha Sharma, Pune, Maharashtra, on Jan. 2, for 'multi-agent artificial intelligence-based system and method for automated evaluation of handwritten examination papers.'
Inventor(s) include Mr. Sujit Sarkar; Mr. Mrityunjoy Panday; and Dr. Neha Sharma.
The application for the patent was published on Feb. 13, under issue no. 07/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses a system for automated evaluation of handwritten examinations. The system ingests scanned answer sheets and performs pixel-domain conditioning to suppress noise while preserving stroke edges, followed by localized thresholding to generate a binary representation. Page skew is detected using aggregated baseline orientation hypotheses and corrected through affine rotation, while peripheral artifacts are removed using connected-component-based border trimming. Handwritten content is then extracted as ordered text lines, tables, and diagrams, and converted into a structured, machine-readable representation with positional metadata. Extracted answer blocks are mapped to examination questions by computing lexical and positional relevance scores and validating pairings through directional agreement thresholds. Validated question-answer pairs are evaluated by multiple coordinated AI processes that assess factual coverage and clarity in accordance with predefined rubrics and reconcile scores through a consensus mechanism. All intermediate evaluations are logged for auditability, and a report generator produces per-student results with scoring breakdowns and feedback for electronic dissemination."
Disclaimer: Curated by HT Syndication.
