論文提出了一種新的圖象分類算法——基于微粒群的圖象分類算法。將此算法和K 均值聚類算法分別應用于MRI 人腦圖象的分類，并進行了比較。實驗結果表明：基于微粒群的圖象分類算法具有較好的全局收斂性，不僅能有效克服K 均值算法易陷入局部極小
值的缺點，而且全局收斂性能優于K 均值算法。
關鍵詞:微粒群算法；K 均值算法；圖象分類
Research of Image Classification Method Based on Particle Swarm Optimization Zhou Xian-cheng （Department of Computer Science and Electronic Engineering Hunan Business College， Changsha Hunan 410205,China） [ABSTRACT]:This paper proposes a new image classification algorithm based on Particle Swarm Optimization.The new image classification algorithm has been applied to MRI images to illustrate its applicability.The experimental results show that the proposed algorithm not only avoids the local optima,but also has greater searching capability than K-means algorithm,yielding promising results.
[Key Words]: Particle swarm optimization; K-means algorithm; Image classification