Interactive Image Segmentation Based on Multi-cue Dynamic Integration
基于多线索动态融合的交互式图像分割

In order to improve the stability and generality of image segmentation, a new method based on the dynamic multi-cue integration is presented in this paper. It fuses color, texture, spatial and edge information by the construction of a conditional random field for image segmentation. Besides, in the course of interactive image segmentation, the proposed method is able to take full advantage of the user labeled information about the foreground and background, and builds a certain standard to measure the reliability of established probability distribution, with which to assemble the energy terms in the conditional random field. This makes the related energy terms fusing dynamically in accordance with the inner property of the image,and improves the segmentation power of the model. The experimental results demonstrate that the proposed method performs steadily and works well on various natural images.