Physical Structure Segmentation with Projection Profile for Mathematic Formulae and Graphics in Academic Paper Images
Tzu-Yuan Chang, Yusuke TAKIGUCHI, Minoru Okada · Proceedings of the International Conference on Document Analysis and Recognition · 2007
Image segmentation is an essential topic in a document processing system, and also is an important initial task for higher level image processing such as recognition or object tracking since recent years. In this paper we propose a method for segmentation and extraction of the regions of mathematical formulae and graphics in an academic paper. First, an academic paper image is transformed into a binary image by Ohtsu's method. Second, the boundary blank spaces of the binary image are removed and then the image is recursively segmented and extracted by vertical and horizontal projection profiles. We apply the proposed method to 150 academic paper images, and show that the method is adaptable, robust and effective for segmentation.