Video Text Extraction from Images for Character Recognition

Text extraction from images and video sequences finds many useful applications in document processing and parts identification in industrial automation. Text extraction may employ binarization or directly process the original image. In the conventional method, grey image binarization processing with a given threshold is employed to extract high intensity video character regions. A video has a complex background with various kinds of intensity so that appropriate thresholds are not always obtained. The proposed method describes the flow of character segmentation from the image in order to make a readable image for optical character recognition. The intensity distribution of characters in an image precisely estimated to extract only character regions from the image. The main objective of this method is to reduce the number of false character regions that may be fed to optical character reader. Incorporating an optical character reader algorithm with the proposed method yields a useful system for text analysis in images