Medical image classification as an important researchtopic both in image processing and biomedical engineering. Theridgelet transform has good directional selective ability to locallyand sparsely in representing the image compared with thetraditional wavelet transform. This paper proposes a novelclassification model for medical image, which is using ridgelettransform and dynamic fuzzy theory. Firstly, the image wasdecomposed by digital ridgelet transform to obtain theapproximation coefficients and detailed coefficients in differentsub-bands with directional parameters. Then the dynamic fuzzytheory was applied to construct a membership function tocalculate coefficients from each sub-bands respectively, and aweight of sub-bands degree was adjust by precision requirement.At last similarity degrees are calculated by coefficients degreeand weight. Medical images were classified by the result sortorder of the degrees effectively.
展开▼
