关键词: 图像分割, 区域生长, 种子, 微粒群优化算法

Abstract: A new region growing algorithm for image segmentation is proposed，which is based on Particle Swarm Optimization algorithm by utilizing particle swarm’s power searching ability to search pixel seeds.Because searching pixel seeds are based on density and the computational load is small，the computing speed of the algorithm can be improved obviously.Compared to traditional RG method，the proposed algorithm can overcome the disadvantages that traditional RG method can’t select seeds automatically and leads to over-segment.The results indicate that the algorithm can segment the image accurately and precisely.

Key words: image segmentation, region growing, seed, particle swarm optimization algorithm