An Automatic Approach for Translating Simple Images into Text Descriptions and Speech for Visually Impaired People

Image processing is a rapidly growing field of research. Images are of different file formats and of different things, places, humans, scientific, astrological and many such. An image is a collection of several pixels arranged in rows and columns. These images are captured, processed and stored for various uses. For common people it is very easy to identify and analyze general images but for the blind and physically disabled people it is difficult. Unfortunately, there is no prior medium or interface for such needy people to communicate with the world. Blind or visually impaired people are usually those people who are neglected by the society, so there is always a need to help such people. Hence, we propose a new technique of converting images into text as well as speech using techniques provided by image processing like pre-processing, image segmentation, edge detection, object detection and speech synthesis. In this paper we first introduce image to text conversion need for blind people and system overview of image to text and speech conversion system. Edge detection plays an important role in this system where Canny edge detection algorithm is used to detect objects from images. Object recognition is done on the basis of color, size, texture and shape of the object.