An image segmentation approach based on histogram analysis utilizing cloud model

The paper focuses on the image segmentation methods based on histogram analysis, and proposes a novel image segmentation approach based on cloud model. Firstly, the paper introduces the basic principles of cloud model. Similar to type-2 fuzzy sets, cloud model considers the uncertainty of membership grades. But it also considers the randomness of them. It is a new kind of uncertain model which is different from type-2 fuzzy sets. Secondly, the proposed image segmentation approach is described. The histogram of image is transformed into discrete quality concepts expressed by cloud models. Based on these quality concepts represented by cloud models, image segmentation is realized by the principle of maximum certainty degree. In the end, the paper compares the proposed method with fuzzy C means (FCM) method and Gaussian mixture model (GMM) method. Experiments demonstrate the effectiveness of the proposed method.