利用Fisher线性判别细化K-均值聚类算法的灰度图像分割

Improving K-means Clustering Using Fisher Criterion in Image Segmentation

聚类是图像分割的一种通用方法,K-均值法聚类图像分割具有一定的自适应性,但结果易受聚类中心和几何形状的影响。本文引入Fisher线性判别对K-均值分割的结果进行不断的迭代改进。实验结果表明该方法可以提高分割的精度和准确度。

Clustering is a common technique for image segmentation, k-mean clustering algorithm is self-adaptable, unsupervised clustering algorithm in image segmentation. But the results are easy to be affected by their clustering centers. Fisher criterion is introduced to iteratively improve the result of image segmentation from k-mean clustering algorithm. The experimental results show that this method can improve accuracy and precision of image segmentation.