基于FCM的分块自适应图像分割方法研究

基于模糊C均值聚类（FCM）的图像分割是应用较为广泛的图像分割方法之一,但是传统的模糊C均值聚类算法都是基于欧氏距离的,对于图像中的噪声是十分敏感的。针对这一局限性,提出一种基于FCM的分块自适应图像分割方法。该方法不仅考虑了噪声不均匀分布对分割结果的影响,而且充分考虑了图像像素的灰度信息和空间信息。通过对含有噪声的自然图像和合成图像的分割试验,我们可以得到,与传统的FCM图像分割算法相比,本文方法能显著提高含有噪声图像的分割质量。

Research of block adaptive image segmentation method based on fuzzy C-mean

Image segmentation based on fuzzy C-means clustering （FCM） is one of the more widespread application of image segmentation methods, but the traditional fuzzy C-means clustering algorithm segmentation of the image is based on Euclidean distance, and very sensitive to the noise. In order to overcome this limitation, a block adaptive FCM image segmentation is proposed. This algorithm not only considers the impact of uneven distribution of the noise, but also takes full account of the image pixel gray-scale and spatial information. Through the segmentation experiments of natural and synthetic images containing noise, we will obtain that compared with the traditional FCM segmentation algorithm, the proposed method can significantly improve the segmentation quality of noisy images.