Text segmentation in mixed-mode images

Block based algorithms have found widespread use in image and video compression. However, popular algorithms such as JPEG, which are very effective in compressing continuous tone images, do not perform well with mixed-mode images which have a substantial text component. With a growing number of applications where such images occur, e.g., color facsimile, digital libraries and educational videos, there are advantages in being able to classify each block as being text or continuous tone. With such a classification, different compression parameters or even algorithms may be employed for the two kinds of data to obtain high compression with minimal loss in visual quality. In this paper we analyze and compare four methods for block classification in mixed mode images, namely variance, absolute-deviation, edge, and DCT based methods. Our evaluation of each scheme is based on the accuracy of segmentation, robustness across different types of images and sensitivity to the threshold used for segmentation. Our results show that DCT based segmentation offers the best accuracy and robustness. Another advantage of DCT is that it is compatible with standards like JPEG, MPEG and H.261