【摘要】 针对传统FCM(模糊C均值)聚类算法及改进算法无法对背景有大片点状、片状斑纹以及字迹模糊的甲骨文字图像进行有效分割的情况,提出了一种基于二进小波变换与FCM聚类算法的甲骨文字图像分割算法.首先,采用二进小波变换模极大值点对甲骨文字图像进行边缘检测;然后,充分利用二进小波变换模极大值中的边缘信息,从而进一步修改FCM聚类算法中的隶属度函数.将实验结果与传统的FCM聚类算法及改进算法进行比较,证明了该算法能更有效地分割甲骨文字图像,具有更高的正确分割率.更多还原

【Abstract】 To solve the problem that the traditional FCM(fuzzy C means)fuzzy clustering algorithm and the improved algorithm cannot effectively separate the JiaGuWen image and there is a large number of dotted or flaky stripes on its background,an image segmentation algorithm based on wavelet transform and FCM fuzzy clustering is proposed. Firstly,the edge information of the JiaGuWen image is detected by using the dyadic wavelet transform modulus maxima,and then the edge information in the maximum value of the binary wavelet transform is used to further modify the membership function in FCM fuzzy clustering. Finally,the experimental results are compared with the traditional fuzzy clustering algorithm and improved algorithm.The results proved that this algorithm can more occurately segment the images of characters.更多还原