1. A new active contour model based on complexity of textures for segmentation of natural image
- Author
-
Baozhu Guo, Zhizhong Fu, Hongsheng Liu, and Xiaofeng Li
- Subjects
Active contour model ,Computer science ,business.industry ,Diffusion filter ,ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION ,Pattern recognition ,02 engineering and technology ,Image segmentation ,01 natural sciences ,Measure (mathematics) ,Image (mathematics) ,Level set ,Feature (computer vision) ,Computer Science::Computer Vision and Pattern Recognition ,0103 physical sciences ,0202 electrical engineering, electronic engineering, information engineering ,020201 artificial intelligence & image processing ,Segmentation ,Artificial intelligence ,010306 general physics ,business - Abstract
Segmentation of natural image with complex textures is a challenge problem for classical Active Contour Models, because it is difficult to find a suitable measure for describing textures that always occupy a big area of natural image. In this paper, a new measure to evaluate texture complexity is proposed, then a feature map is created based on this measure. To approximate two phase image, diffusion filter is used to smooth the feature map. Finally, we proposed a new Active Contour Model based on the feature map. This new Active Contour Model can be used to segment natural image. Results of experiments on some typical natural images show the proposed model suffer from less over and wrong segmentation than state-of-the-art models.
- Published
- 2017