MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202531134092 A) filed by Kalinga Institute Of Industrial Technology, Bhubaneswar, Orissa, on Dec. 31, 2025, for 'hybrid deep learning system for optical character recognition and classification of medical prescriptions.'
Inventor(s) include Dey, Raghunath; and Piri, Jayashree.
The application for the patent was published on Jan. 23, under issue no. 04/2026.
According to the abstract released by the Intellectual Property India: "The present invention provides a hybrid deep learning system (100) for optical character recognition and classification of medical prescriptions comprising an image acquisition module (102) for capturing prescription images, a preprocessing unit (104) operatively connected to perform grayscale conversion, adaptive thresholding, geometric correction, and noise reduction, a convolutional neural network encoder (106) for extracting spatial features using multi-scale kernels, a bidirectional long short-term memory decoder (108) with attention mechanism module (110) for processing temporal dependencies, a sequence-to-sequence architecture (112) enabling end-to-end text recognition, a vision transformer classifier (116) for identifying medical entities, a YOLO-based text region detection module for localizing prescription fields, a post-processing unit (118) performing spell correction, regex-based extraction, fuzzy matching, and market basket analysis, and an output interface (120) generating structured outputs in JSON, CSV, and database formats, achieving 91% recognition accuracy across diverse handwriting styles."
Disclaimer: Curated by HT Syndication.
