Survey of Different Techniques for Text Detection from Natural Images

Abstract

Texture analysis can provide very useful and vital information for content-based image analysis. Text recognition and analysis includes many applications such as: license plate recognition, sign detection as well translation, helping tourists and blind persons to understanding the environment, drawing attention of a driver, content-based image search and so on. Locating text in case of variation in style, colour, as well as complex image background makes text reading from images more challenging. in this paper the various techniques available for detecting and recognizing text are explained, finally a hybrid approach using segmentation explained which can improve the qualitative texture analysis among other techniques.