A deformable local image descriptor Export

@inproceedings{cheng08, abstract = {This paper presents a novel local image descriptor that is robust to general image deformations. A limitation with traditional image descriptors is that they use a single support region for each interest point. For general image deformations, the amount of deformation for each location varies and is unpredictable such that it is difficult to choose the best scale of the support region. To overcome this difficulty, we propose to use multiple support regions of different sizes surrounding an interest point. A feature vector is computed for each support region, and the concatenation of these feature vectors forms the descriptor for this interest point. Furthermore, we propose a new similarity measure model, local-to-global similarity (LGS) model, for point matching that takes advantage of the multi-size support regions. Each support region acts as a dasiaweakpsila classifier and the weights of these classifiers are learned in an unsupervised manner. The proposed approach is evaluated on a number of images with real and synthetic deformations. The experiment results show that our method outperforms existing techniques under different deformations.}, author = {Cheng, Hong and Liu, Zicheng and Zheng, Nanning and Yang, Jie}, booktitle = {Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on}, citeulike-article-id = {4947629}, citeulike-linkout-0 = {http://dx.doi.org/10.1109/CVPR.2008.4587378}, citeulike-linkout-1 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4587378}, doi = {10.1109/CVPR.2008.4587378}, journal = {Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on}, keywords = {descriptors, features, matching, onur2009}, pages = {1--8}, posted-at = {2009-06-24 21:33:04}, priority = {2}, title = {A deformable local image descriptor}, url = {http://dx.doi.org/10.1109/CVPR.2008.4587378}, year = {2008} }

TY - CONF ID - cheng08 L3 - citeulike-article-id:4947629 N2 - This paper presents a novel local image descriptor that is robust to general image deformations. A limitation with traditional image descriptors is that they use a single support region for each interest point. For general image deformations, the amount of deformation for each location varies and is unpredictable such that it is difficult to choose the best scale of the support region. To overcome this difficulty, we propose to use multiple support regions of different sizes surrounding an interest point. A feature vector is computed for each support region, and the concatenation of these feature vectors forms the descriptor for this interest point. Furthermore, we propose a new similarity measure model, local-to-global similarity (LGS) model, for point matching that takes advantage of the multi-size support regions. Each support region acts as a dasiaweakpsila classifier and the weights of these classifiers are learned in an unsupervised manner. The proposed approach is evaluated on a number of images with real and synthetic deformations. The experiment results show that our method outperforms existing techniques under different deformations. JF - Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on EP - 8 TI - A deformable local image descriptor T2 - Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on SP - 1 KW - descriptors KW - features KW - matching KW - onur2009 AU - Cheng, Hong AU - Liu, Zicheng AU - Zheng, Nanning AU - Yang, Jie PY - 2008/// UR - http://dx.doi.org/10.1109/CVPR.2008.4587378 ER -