Image segmentation based on two-dimensional cloud model

Abstract

There are many uncertainties in image segmentation, which should be handled by theories and methods with uncertainty. Cloud model is a kind of effective method to handle uncertainty, which considers fuzziness, randomness and the connection of them. Image segmentation based on one-dimensional cloud model processes the one-dimensional grayscale histogram, and segments images using one-dimensional cloud transformation and synthesis, but does not consider spatial information of images. The paper proposes a method based on two-dimensional cloud model, which processes two-dimensional grayscale histogram, and segments images using two-dimensional cloud transformation and synthesis. The proposed method considers spatial information of images, and has good performance suppressing noise. Results of many experiments indicate that the proposed method obtains better effect than those of traditional methods, and it is feasible and effective.