Third Eye for Visually Impaired using Artificial Intelligence

Abstract

This paper presents a comprehensive solution utilizing Raspberry Pi technology to assist visually impaired individuals in everyday tasks. The proposed system integrates three key functionalities: family member recognition, real-time object detection, and optical character recognition (OCR). Firstly, the system employs Haarcascade and LBPH algorithms for family member recognition. By detecting a maximum of two faces, it provides audio-based feedback, allowing the user to identify individuals standing in front of them. This proof-of-concept level solution aims to enhance social interactions and facilitate familiarity in dynamic environments. Secondly, real-time object detection functionality enables the system to recognize various objects presented before the camera. Utilizing voice-based feedback, the system provides instant identification and description of objects, aiding visually impaired users in navigating their surroundings and understanding their immediate environment. Lastly, the system incorporates OCR technology to assist users in reading printed text. This proof-of-concept level solution aims to empower visually impaired individuals by converting printed text into audible information, thereby facilitating access to written content in various contexts. This multi-solution approach leverages Raspberry Pi technology to address different aspects of accessibility for visually impaired individuals, aiming to improve their independence, social interaction, and access to information in everyday life