Harris corner detection based leaf image segmentation for ancient Chinese books

In the image digitization process of ancient Chinese books, sometimes it is needed to extract the leaf from the original image and separate the left half and the right half. This paper proposes a leaf segmentation method based on Harris corner detection to solve this problem. Firstly, the original image is preprocessed to recognize the contour information, and the minimum bounding rectangle is determined by the contour. Then, Harris corner detection is used. In order to recognize the edge of the leaf, we traverse the corner points in the special area and save those points which meet the threshold requirement. At last, the edges of left half leaf and right half leaf are determined, and the segmentation is completed. The experimental results show that the proposed method achieves good performance in a class of leaf image segmentation.