关键词: 图像分割, 聚类, 人工免疫模型, 云模型

Abstract: Artificial immune model is introduced in searching clustering centers，that avoids being trapped in local optimal solution.On one hand immunological character of antigens is polyvalent；on the other hand high False Accept Rate（FAR） is produced，because the antigens which are neither recognized by recognizers nor self ones are decided as self ones or non-self ones identically.So cloud model is applied in antigen-antibody recognition，which increases rondomness and vagueness in antigen division and improves image segmentation quality.

Key words: image segmentation, clustering, artificial immune model, cloud model