An Integrated OCR-Based Assistive System for Visually Impaired Individuals with Enhanced Accessibility

Mainstream technologies for assisting specially-abled individuals have evolved in both printed and digital mediums. This research paper presents a study on a system designed to assist specially-abled individuals using Optical Character Recognition (OCR). It also explores modern-day solutions for real-time data accessibility. The OCR system is integrated with a wide range of hardware to enhance accessibility and convenience. The hardware includes keyboards, displays, buzzers, controllers, actuators, and more. For real-time data access, a web server is provided for manual data input, which is then processed and recognized using the software. The input data can be digital, manually typed, or in the form of various file types. Additionally, a webcam is set up to capture and process data from the surroundings for recognition. The software extends its functionality to handwritten notes and other forms of data. It can differentiate between numerals, alphabets, and symbols. The recognized data is then translated into the required Braille format, specifically arrays corresponding to each letter. The translated data is subsequently transmitted to the hardware for appropriate feedback. This research paper also includes a comparative analysis of three widely recognized OCR models—EasyOCR, Pytesseract, and SuryaOCR. The analysis evaluates various performance aspects, including speed, processing time, accuracy, complexity, dependencies, error rate, and error-handling capacity.