Texture classification using reduced set of nonsubsampled contourlet transform features

Texture based classification is an important approach for effective classification of images. In this work, a non-subsampled contourlet transform is employed to extract the directional frequency information followed by the statistical moment extraction where, zernike moments are used as texture descriptors. The main advantage of this approach is that it helps in reducing the dimensionality contourlet coefficients. from the experiments conducted in this work, it has been observed that combining non-subsampled contourlet transform and zernike moments produces good image representative capability. Moreover, nearest neighbour classifier is used in this work as classifier. For the experimental stud, brodatz database of textures is used. From the experimental results, it has been observed that non-subsampled contourlet transform combined with zernike moments achieve greater performance than the other well-known models.