Image Segmentation Based On Ridgelet Bi-frame System

Ridgelet Bi-frame System is a new image representation system based on ridgelet, It inherit the core thinking of the ridgelet that is turn the "line of bizarre" of a image into a "point of singularity", and then to deal with these points of singularity with wavelet. It overcomes the disadvantages of the finite ridgelet for image including "wraping" phenomenon, reconstruction inaccurate, can only be used for the prime size of image.A image segmentation method based on gray level co-occurrence matrix and Ridgelet bi-frame system is presented. The method uses the adaptive size of the window, it combines the features of Gray level co-occurrence matrix and Ridgelet Bi-frame System, then be clustered with fuzzy C-means (FCM). Compared with the wavelet segmentation method and the finite ridgelet segmentation method, It improve the results well.For the SAR images, a image segmentation method based on the watershed algorithm and Ridgelet Bi-frame System is presented, this method uses the advantages of the watershed algorithm on the edge positioning, then be combined with Ridgelet Bi-frame System segmentation results for the final result, experimental results show that the method can effectively improve the edge accuracy and the regional consistency.