ROBUST DETECTION OF TEXT IN NATURAL SCENE IMAGES

Abstract

Detection of text and identification in natural scene images has applications in computer vision systems such as license plate detection, automatic street sign translation, image retrieval and help for visually challenged people. The text images has complex background, blur, occluded text, different font-styles, and noises in image and variation in illumination. Hence, scene text recognition puts forth challenges in computer vision. Hence, a potent method based on Maximally Stable External Regions (MSER) has been used as described in this paper. Here, the text characters are clustered, separating them from high probable non-text characters with the help of text categorizer. The algorithm is then verified by testing it on images based on the predefined rules.