MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202621057374 A) filed by Dr. D. Y. Patil Institute Of Technology, Pimpri, Pune on May 06, 2026, for Ai-Powered Blind Reading.
Inventors include Mrs. Sunita Patil; Mr. Aaryan Pawar; Mr. Pranit Pawar; Mr. Aditya Kulkarni; Mr. Vedant Gaikwad; Mr. Sharad Adsure; and Dr. Rachana Somkunwar.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: This article presents a hybrid AI-enabled assistive system meant to help the visually impaired by translating visual content into audio descriptions. The system combines many OCR tools, image-description models, and modern text-to-speech technology to provide accurate, real-time accessibility support. It uses PaddleOCR for offline recognition, TrOCR for text extraction, and Google Vision API for cloud processing. This process allows the system to switch between OCR methods based on confidence levels, image difficulty, and network availability. When readable text isn’t present, models like BLIP-2 can generate meaningful descriptions of the image. This text is then converted into natural speech using Google WaveNet. The hybrid process guarantees optimum accuracy, efficiency, and flexibility in varying environments. This document highlights the system architecture, methodology, implementation, and assessment, demonstrating its efficiency in practical usage.
Disclaimer: Curated by HT Syndication.
