A Graph-Based Method for Image Segmentation in Cross-Region

Cross-region in scanned image is a complex one in line recognizing process. Using graph-based model, the geometric properties such as cross-attribute can be extracted from image. Then they can be divided into several categories. Based on the topological and adjacent properties of the graph, the principle of image segmentation in cross-region is proposed. The proposed method has advantages over individual pixels due to its intrinsic behaviour, retrieval of colony characteristic of related pixels, shape independence, and ease of implementation.