Text segmentation and recognition in natural scene images using GLCM and FOS feature descriptor

Abstract

In the real world, Text plays major role in the scenery images and newlinevideo frames, understanding that text carries important information in day to newlineday life. Gradually, the number of users of moveable camera, video recorders newlineand camera phone are increasing and this direct to maximum development in newlinemultimedia information. Researchers were found that human pay more newlineattention to the text rather than the object in the images. The main focus of the newlineresearcher is to find the best way of finding and extracting the different types newlineof text in scenery images or in video frames having complex backgrounds. One newlinecan say that it is extending the application of optical character recognition newlinesystem into broader area and also service people for additional understanding newlinethe machine of the text detection and recognition. Actually text detection is the newlinearea in which one tries to find the text area in the scenery images and video newlineframes by finding precise borders of text lines through various approaches. newlineThis proposed work plays a major role in detecting the text region to newlinehelp the aged persons when travel long distance through bus, to see the number newlineboard of the bus. For image acquisition, get a publically available database newlineimages. MSRA-TD500 database is a text image database. The proposed system newlineuses YUV color conversion technique to improve the quality of images, avoid newlinethe unnecessary distortions in the image and then Y channel is converted to newlinegray scale image using gray scale converter, which contains the intensity values newlineranges from 0 to 255. newline newline