Image segmentation application based on the normal cloud model

In image segmentation, one important problem is the indeterminacy of pixel value. Based on fuzzy and probability theory, Cloud Model can better solve this same concept problem in segmentation. Methods based on Normal Cloud Model have been proposed here for improving the segmentation performance. Single-Kernel Extraction (SKE) and Multi-Kernel Extraction (MKE) methods have been introduced which are used for determining expected value of the Cloud Model during Cloud Transformation. In addition, a Maximum-Similarity Concept Promotion (MSCP) strategy based on Minimum-Distance Concept Promotion (MDCP) has been proposed and its theoretical validity has also been introduced. Image segmentation experiment results show that our algorithms get a strong adaptability and better image segmentation effect evaluation coefficient.