Figure 4 (7): Example of cascading decomposition using Tree- structured wavelet transform. Texture Features Vector: The first 5 largest energy values of leaf nodes in the energy map were selected as features.
Related Figures (6)
![Figure (5): One level of 2-D multi wavelet decomposition of a 2-D image decomposition, 16-subband intermediate image will be as shown in figure (5).The next decomposition will be in the "low- low pass" sub-matrix, the shaded part in figure (6). According to advance process, an L-level decomposition of a 2-D image will produce 4(3L+1) sub-bands [12].](https://figures.academia-assets.com/46525533/figure_003.jpg)
![Image decomposition: As in TSWT, DITCWT depends on calculating the energy in the sub-bands images to continue the cascading decomposition. But, DTCWT instead of decomposing the image to 4 sub-band images, each texture image was decomposed to 8 sub images (2 low pass, and 6 band pass). The next level of decomposition was to that sub-images which had a significant value of energy -using same equation (3) - for extra 8 sub images. Figure (9) illustrates an example for one texture image [15]. Figure (9): Decomposition example for a texture image using DITCWT.](https://figures.academia-assets.com/46525533/figure_006.jpg)
Figure 4 (7): Example of cascading decomposition using Tree- structured wavelet transform. Texture Features Vector: The first 5 largest energy values of leaf nodes in the energy map were selected as features.