Camera-Based Document Image Layout Analysis System On Android

With the development and popularization of mobile terminal, more and more people are wild about looking up information and doing mobile office with portable devices. As the objects of information retrieval and information processing, Images with a large amount of information and extensive content are widely used in all kinds of information processing applications. The image processing has become the main work, especially the text and graphic segmentation and extraction become the key of image processing.Document image layout analysis means to layout segmentation and distinguish the text and graphic area automatically according to different feature extraction and matching. Nowadays, layout analysis research is mainly for the realization in the PC platform. Because uneven illumination of filming and image noise interference, at the same time due to the complexity of algorithm and resource consumption, which does not apply to the mobile terminal.This article analyzes the problems that exist in the development of image processing program on Android platform, and the current image processing, binarization, run smooth, connected component analysis methods were studied. So we make optimization and improvement to against the problems, which can reduce the consumption of resources. This article uses a method about background estimation, it can make a better effect on binary images which captured under non-uniform illumination environment. Then we propose an adaptive run smooth method which make the image form connected domain by block, so that we can layout segmentation the document image. Finally, analyzing and comparing the characteristic of the segmented regions, aiming to design an Android platform layout analysis system so as to distinguish the text and graphic area on a document image.