An Efficient Method for Text Detection and Recognition in Still Images

Abstract

Text Detection and recognition in general have quite a lot of relevant application milestones in various fields such as unmanned drone, defence vehicle, sign board detection, security surveillance, and even for unmanned driving. Text detection from blurred images due to environment weather is very challenging. The main aim of this paper is to identify and recognize the text in still images. Results demonstrated that the new adaptive algorithm could efficiently recognize the text even if the image is corrupted by high noise density. Also implementation of the proposed method is easy and less complex when compared to other methods. This paper shows how to split regions containing text in an image and how to detect the text. The automated text detection algorithm in this paper uses MSER regions and optical character recognition (OCR) is used to recognize the text. Here system detects the text and finds the connected regions, chain them together in their relative position. The result of multiple segmentation hypotheses are post-processed by a connected component analysis algorithm. This algorithm is applied to several sample images and proved that this method is more efficient.