An efficient scheme for detecting MSER regions and text localization in video

In the present period of IT and correspondence innovation, use of video based data is expanding enormously. Efficient algorithms are very highly demanded Detection of scene text and caption text in the video in the area of image processing. Every single video contains some imperative data which should be separated without losing any data from the video, much consideration is currently given to the video processing innovation. In this paper, A method has been proposed to efficiently extract the frames from the videos based on color moments and then detecting Maximally Stable Extremely Regions (MSER). This method is done only on key frames which will work until text localization. This method is helped in reducing the processing time of algorithm for text localization.