Word Based Text Extraction Algorithm Implementation in Wearable Assistive Device for the Blind

Authors:

Abstract
The absence of proper Braille based resources for all kind of documents has adversely affected the life of Visually Impaired people. This paper presents a wearable assistive device for the blind which converts the text into acoustic output enabling the user to read any sort of text. The focus of this paper is the standalone Raspberry Pi based system with finger mounted camera that can help the visually impaired people in word based reading of the textual data pointed to by the finger. The system consists of a webcam that captures images which are enhanced. Following this the word pointed by the finger is extracted using a novel methodology and given to an Optical Character Recognition (OCR) engine. Subsequently, the textual output is given to a Text to Speech (TTS) converter to obtain audio via earphones.