关键词: 图像分割, 性能评估, 灰色聚类

Abstract: In view of the performance analyses of image segmentation, a hierarchical synthetic evaluation method is proposed based on the grey clustering. The method obtains the appropriate albinism function by target characteristic, then the scattered information produces cluster matrix. The new method can distinguish the performances from different segmentation algorithms easily. The experiments show that the method has higher sensitivity and lower singularity than other methods, and can be more effectively used to analyze the performance of segmentation algorithms.

Key words: image segmentation, performance evaluation, grey clustering