Description

Text Detection in Images Using Python

Text detection in images or videos is an important step to achieve multimedia content retrieval. In this paper, an efficient algorithm which can automatically detect, localize and extract horizontally aligned text in images with complex backgrounds is presented. The proposed approach is based on the application of a colour reduction technique, a method for edge detection, and the localization of text regions using projection profile analyses and geometrical properties. The text contains vital and useful information which is embedded in various types of documents and natural scene. The extraction of text from a natural image is a challenging task. The text is detected and is extracted in this way that it readable by another person without any difficulty. In this paper we proposed a fast text localization method in which sobel edge detection and clustering method is used to localize all the possible edges in the image.

Demo Video