Year: 2025
Type: article
Abstract: Visually impaired users find it extremely hard to navigate environments, recognize objects, and read printed let- ters. Existing assistive technologies are marred with high costs, clunky interfaces, and reliance on internet connectivity. This paper presents a mobile application that solves these problems utilizing real-time computer vision and AI. The application integrates YOLO for object recognition, Tesseract-based OCR for text recognition, and TTS synthesis for multimodal, hands-free output. Advances include voice-controlled automatic switching of modes, offline operation with TensorFlow Lite power, and cross-platform compatibility with a React Native framework. Performance experiments demonstrate 95% text recognition accuracy under favorable lighting and 89% object detection ac- curacy with execution time at 28ms on average—much improved over existing solutions like Google Lookout. The application is immune to internet connectivity, reduces cognitive load through voice-controlled interaction, and is less expensive with stock smartphone hardware. The comparative study demonstrates advantages of affordability, accuracy, and offline capability, and the solution is a practical tool to render visually impaired users independent. Future development involves expanding the number of detectable objects and depth sensing to minimize optical illusion error.
Authors: Anisha Raskar, Ankita Dhanbhar, Harshada Mhaske, Khushal Patil, Prof. Vijay Sonawane
Language: English
Subfield: Occupational Therapy
Field: Health Professions
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W4410129281