Abstract: A new image texture segmentation method is presented based on the Krawtchouk moments and support vector machine (SVM). The Krawtchouk moments in small local windows of each pixel in the image are computed and a nonlinear transducer is used to map the moments to texture features. The feature vector is then input to SVM for classification. Compared with the segmentation results based on the Zernike moment, the proposed method can produce better results.

Key words: Krawtchouk moment, SVM, image segmentation, texture segmentation, texture analysis, image analysis