An effective processing method for biomedical images and the Fuzzy C-mean (FCM) algorithm based on the wavelet transform are investigated.By using hierarchical wavelet decomposition, an original image could be decompo...An effective processing method for biomedical images and the Fuzzy C-mean (FCM) algorithm based on the wavelet transform are investigated.By using hierarchical wavelet decomposition, an original image could be decomposed into one lower image and several detail images. The segmentation started at the lowest resolution with the FCM clustering algorithm and the texture feature extracted from various sub-bands. With the improvement of the FCM algorithm, FCM alternation frequency was decreased and the accuracy of segmentation was advanced.展开更多
为了提高基于拍摄方式的文档图像的二值化效果,降低光学字符识别(optical character recognition,OCR)系统的文字识别错误率,提出了一种全局阈值与局部阈值相结合的二值化算法——VFCM。该算法使用最大方差比方法产生全局阈值,使用FCM(F...为了提高基于拍摄方式的文档图像的二值化效果,降低光学字符识别(optical character recognition,OCR)系统的文字识别错误率,提出了一种全局阈值与局部阈值相结合的二值化算法——VFCM。该算法使用最大方差比方法产生全局阈值,使用FCM(FuzzyC-Means)聚类方法产生局部阈值。这两种方法的结合能够较好地保留字符的笔画细节,并能有效地消除伪影。实验结果表明,该算法可以取得比较好的二值化效果,并能带来OCR系统识别率的有效提高。展开更多
文摘An effective processing method for biomedical images and the Fuzzy C-mean (FCM) algorithm based on the wavelet transform are investigated.By using hierarchical wavelet decomposition, an original image could be decomposed into one lower image and several detail images. The segmentation started at the lowest resolution with the FCM clustering algorithm and the texture feature extracted from various sub-bands. With the improvement of the FCM algorithm, FCM alternation frequency was decreased and the accuracy of segmentation was advanced.
文摘为了提高基于拍摄方式的文档图像的二值化效果,降低光学字符识别(optical character recognition,OCR)系统的文字识别错误率,提出了一种全局阈值与局部阈值相结合的二值化算法——VFCM。该算法使用最大方差比方法产生全局阈值,使用FCM(FuzzyC-Means)聚类方法产生局部阈值。这两种方法的结合能够较好地保留字符的笔画细节,并能有效地消除伪影。实验结果表明,该算法可以取得比较好的二值化效果,并能带来OCR系统识别率的有效提高。
