Abstract:

Therefore, after analyzing the importance of frequency domain phase information and textural information in characterizing image features, a novel SVM image segmentation method is proposed using phase consistency and textural features. The new method combines phase consistency statistic characteristics, textural features and graylevel characteristics into a training eigenvector and segments image with SVM classification technique. Compared with the traditional method, the statistical eigenvectors extracted by the new method can reflect details of the edges of image and textural information effectively. The experimental results show that the new method is more effective than the traditional method for SVM image segmentation, especially in the situation where there is low edge contrast and rich textural information in the image's target area. 

Key words: image segmentation, phase consistency, textural features, SVM