Feature hierarchies for object classification Export

@proceedings{Epshtein2005, abstract = {The paper describes a method for automatically extracting informative feature hierarchies for object classification, and shows the advantage of the features constructed hierarchically over previous methods. The extraction process proceeds in a top-down manner: informative top-level fragments are extracted first, and by a repeated application of the same feature extraction process the classification fragments are broken down successively into their own optimal components. The hierarchical decomposition terminates with atomic features that cannot be usefully decomposed into simpler features. The entire hierarchy, the different features and sub-features, and their optimal parameters, are learned during a training phase using training examples. Experimental comparisons show that these feature hierarchies are significantly more informative and better for classification compared with similar nonhierarchical features as well as previous methods for using feature hierarchies.}, author = {Epshtein, B. and Uliman, S.}, booktitle = {Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on}, citeulike-article-id = {4493648}, citeulike-linkout-0 = {http://dx.doi.org/10.1109/ICCV.2005.98}, citeulike-linkout-1 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1541260}, doi = {10.1109/ICCV.2005.98}, journal = {Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on}, keywords = {dfg0509}, pages = {220--227 Vol. 1}, posted-at = {2009-05-08 13:36:24}, priority = {2}, title = {Feature hierarchies for object classification}, url = {http://dx.doi.org/10.1109/ICCV.2005.98}, volume = {1}, year = {2005} }

TY - CONF ID - Epshtein2005 L3 - citeulike-article-id:4493648 N2 - The paper describes a method for automatically extracting informative feature hierarchies for object classification, and shows the advantage of the features constructed hierarchically over previous methods. The extraction process proceeds in a top-down manner: informative top-level fragments are extracted first, and by a repeated application of the same feature extraction process the classification fragments are broken down successively into their own optimal components. The hierarchical decomposition terminates with atomic features that cannot be usefully decomposed into simpler features. The entire hierarchy, the different features and sub-features, and their optimal parameters, are learned during a training phase using training examples. Experimental comparisons show that these feature hierarchies are significantly more informative and better for classification compared with similar nonhierarchical features as well as previous methods for using feature hierarchies. JF - Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on EP - 227 Vol. 1 TI - Feature hierarchies for object classification VL - 1 T2 - Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on SP - 220 KW - dfg0509 AU - Epshtein, B AU - Uliman, S PY - 2005/// UR - http://dx.doi.org/10.1109/ICCV.2005.98 ER -