Texture classification: are filter banks necessary? Export

@inproceedings{Varma:2003:DoWeNeedFilter, abstract = {We question the role that large scale filter banks have traditionally played in texture classification. It is demonstrated that textures can be classified using the joint distribution of intensity values over extremely compact neighborhoods (starting from as small as 3 \× 3 pixels square), and that this outperforms classification using filter banks with large support. We develop a novel texton based representation, which is suited to modeling this joint neighborhood distribution for MRFs. The representation is learnt from training images, and then used to classify novel images (with unknown viewpoint and lighting) into texture classes. The power of the method is demonstrated by classifying over 2800 images of all 61 textures present in the Columbia-Utrecht database. The classification performance surpasses that of recent state-of-the-art filter bank based classifiers such as Leung \& Malik, Cula \& Dana, and Varma \& Zisserman.}, author = {Varma, M. and Zisserman, A.}, booktitle = {Computer Vision and Pattern Recognition, 2003. Proceedings. 2003 IEEE Computer Society Conference on}, citeulike-article-id = {1603551}, citeulike-linkout-0 = {http://dx.doi.org/10.1109/CVPR.2003.1211534}, citeulike-linkout-1 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1211534}, day = {15}, doi = {10.1109/CVPR.2003.1211534}, journal = {Computer Vision and Pattern Recognition, 2003. Proceedings. 2003 IEEE Computer Society Conference on}, keywords = {descriptor, evaluation, filters, texture}, month = {July}, pages = {II--691-8 vol.2}, posted-at = {2007-10-01 12:00:44}, priority = {5}, title = {Texture classification: are filter banks necessary?}, url = {http://dx.doi.org/10.1109/CVPR.2003.1211534}, volume = {2}, year = {2003} }

TY - CONF ID - Varma:2003:DoWeNeedFilter L3 - citeulike-article-id:1603551 N2 - We question the role that large scale filter banks have traditionally played in texture classification. It is demonstrated that textures can be classified using the joint distribution of intensity values over extremely compact neighborhoods (starting from as small as 3 × 3 pixels square), and that this outperforms classification using filter banks with large support. We develop a novel texton based representation, which is suited to modeling this joint neighborhood distribution for MRFs. The representation is learnt from training images, and then used to classify novel images (with unknown viewpoint and lighting) into texture classes. The power of the method is demonstrated by classifying over 2800 images of all 61 textures present in the Columbia-Utrecht database. The classification performance surpasses that of recent state-of-the-art filter bank based classifiers such as Leung & Malik, Cula & Dana, and Varma & Zisserman. JF - Computer Vision and Pattern Recognition, 2003. Proceedings. 2003 IEEE Computer Society Conference on EP - 691-8 vol.2 TI - Texture classification: are filter banks necessary? VL - 2 T2 - Computer Vision and Pattern Recognition, 2003. Proceedings. 2003 IEEE Computer Society Conference on SP - II KW - descriptor KW - evaluation KW - filters KW - texture AU - Varma, M AU - Zisserman, A PY - 2003/07/15/ UR - http://dx.doi.org/10.1109/CVPR.2003.1211534 ER -