Enhancing Accessibility: A Real-Time Mobile Application for Visual Narration and Text Recognition for the Visually Impaired

People with visual impairments frequently struggle to comprehend their environment. This study proposes a mobile application designed to provide real-time descriptions of activities occurring around visually impaired individuals. By capturing live video through the mobile camera, the app generates a descriptive narrative of the video, which is conveyed to the user via audio output, enabling a more accessible experience of their environment. The system employs OpenCV for capturing visual data, while Tesseract OCR is used to identify and recognize text from the documents. The Google Text-to-Speech API converts this text into speech, using libraries like Pydub and Librosa for additional audio processing. The proposed system hence results in advancement of accessibility to surroundings, through automatic event localization and text-based descriptions of complex visual data.