混合编码方式的图像聚类算法

Image cluster algorithm of hybrid encoding method

基于群体智能优化算法的图像聚类分析,大多数都采用单一的编码方式,使搜索空间过于局限,算法很容易陷入局部最优,为了解决这个问题,提出一种混合编码方式的图像聚类分析算法(HEICA)。该算法构建一种基于图像聚类的混合编码模型,在扩大搜索空间范围的同时,与改进的雨林算法(IRFA)和量子粒子群算法(QPSO)相结合,提高全局搜索能力。在仿真实验中,采用4组数据集对算法进行聚类有效性测试,并将其与4种常用的聚类算法进行对比,实验结果表明该算法具有较强的全局搜索能力,稳定性高、聚类效果好。

In the clustering analysis based on swarm intelligence optimization algorithm,the most of encoding method only used single form,and this method might be limit range of search space,the algorithm was easy to fall into local optimum.In order to solve this problem,image clustering algorithm of hybrid encoding（HEICA） was proposed.Firstly,a hybrid encoding model based on image clustering was established,this method could expand the scope of the search space.Meanwhile,it was combined with two optimization algorithms which improved rain forest algorithm（IRFA） and quantum particle swarm optimization（QPSO）,this method could improve the global search capability.In the simulation experiment,it was carried out to illustrate the performance of the proposed method based on four datasets.Compared with results form four measured cluster algorithm.The experimental results show that the algorithm has strong global search capability,high stability and clustering effect.