Enhancing Accessibility: An AI-Powered OCR System for Real-Time Text and Object Recognition to Assist Visually Impaired Users

Current OCR (Optical Character Recognition) solutions for the visually impaired are limited in adaptability, real-time identification, and multimodal feedback, making them less efficient. This research paper presents an enhanced AI-powered OCR system that intends to bridge the mentioned limitations through the use of adaptive learning, real-time surrounding text identification, and interactive personalized functions. Supporting navigation and safety purposes, the system is also complemented through object identification via the use of the COCO dataset to identify external objects like traffic signs. The system provides multimodal feedback, multiscript support, text reduction, and integration with assistive technology such as wearables and personalized GPS. This new system has the capability to dramatically improve the life quality and learning outcomes of the visually impaired.