Image Segmentation Based on Similarity of Hierarchical Regions
Hongtao He, Bo Peng, Junzhou Chen · 2019
This paper addresses the problem of hierarchical selection in multi-level image segmentation. In this paper, we propose a new framework: First, we use the hidden regional properties of multi-level image segmentation, selecting the finer object segment regions through the comparison between hierarchies without the aid of ground truth or the users prior knowledge of the object, using the saliency of the object for auxiliary analysis, and optimizing its combination. In the next step, we regard the obtained object segment regions as the seed regions, and Graph Cut is performed on the bottom layer of the multi-level image segmentation result to obtain the segmentation result. The experimental results show that our framework is even better than those of the hierarchical segmentation selected in the multi-level image segmentation results. The proposed framework can be applied to solve the hierarchical selection problem in other cases.