Text to Audio Conversion System with Efficient Portable Camera for Visually Impaired People

Abstract

It's a lot of work to accurately recognize text from a picture and turn it into audio. Our
suggested system analyzes and compares the many methods utilized for text extraction from color
photos, such as text detection and identification. Multiple subtasks, such as text detection, text
localization, text classification, text segmentation, and text recognition, make up this overall assignment.
Transcribing the data included in these photos into English will make using the data more efficient and
convenient. Text extraction is the technique of removing text from a picture. Information retrieval,
keyword searching, editing, documenting, archiving, and reporting all use text extraction in one way or
another. However, poor picture contrast and complex backgrounds, as well as differences in text size,
orientation, style, and alignment, make this a very tough and demanding subject to solve. Character
properties of fonts used in texts and picture quality provide further difficulties. Because of these
obstacles, computers can't read the characters perfectly and identify them. Using Python 2.7.15, Optical
Character Recognition (OCR) technology, and an audio converter, we have created a system that can
extract text from photos and play it back.