基于水平集的彩色图像分割方法

Level Set Based Color Image Segmentation

实现了一种新的彩色图像融合分割方法。首先,通过对彩色图像进行YCbCr变换,提取其颜色特征。进一步,对亮度图像进行Gabor变换,提取其纹理特征。同时,为了消除特征向量的冗余,引入PCA变换对高维特征向量进行压缩。然后,采取向量C-V模型对提取的特征向量进行分割。最后,对分割图像进一步后处理,消除小的区域。真实彩色图像实验结果证明了方法的有效性。

A new method about color image segmentation is proposed.Firstly,the color image is transformed to YCbCr color space to extract color features,and the texture features are extracted from luminance image by using Gabor Filters.Moreover,PCA transform is used to eliminate the redundance of feature vector.Secondly,the vector C-V model is used to segment the feature vector.At last,a further processing is used to eliminat the small regions.Experimental results demonstrate the effectiveness of the method.