基于改进FCM算法的彩色图像破损区域提取

现有的大部分图像修复技术需要人工确定待修复区域。结合改进的FCM算法提出了一种自适应提取彩色图像破损区域的方法。该方法可以自适应获取彩色图像初始聚类数目,并采用交叉熵距离测度进行FCM聚类,同时利用颜色和纹理特征向量对彩色图像进行分割,进而提取破损区域。实验结果表明,该方法不仅能够有效提取图像的破损区域,而且算法的普适度也得到了相应提高。与传统的FCM算法相比,本文方法对彩色图像的分割更易于实现,分割效果令人满意。

Extract Damaged Areas of Color Images with the Improved FCM Algorithm

Considering the most of the existing image restoration techniques,the repaired area need to be manually determined,a new method combined with the improved FCM algorithm is proposed to extract the damaged areas of color image.This method can be adaptive to obtain the number of initial clusters of color images,and use the theory of cross-entropy distance to operate procedure of FCM algorithm,then use the color and texture features to build a vectors for the color image segmentation,and extract the damaged area in the last.Experimental results show that this method can not only extract the damaged areas,but also improve the algorithm’s general moderate.Compared with the traditional FCM algorithm,our algorithm is easier to achieve color image segmentation and segmentation results are satisfactory.