Enhanced biologically inspired model Export

@inproceedings{citeulike:5816896, abstract = {It has been demonstrated by Serre et al. that the biologically inspired model (BIM) is effective for object recognition. It outperforms many state-of-the-art methods in challenging databases. However, BIM has the following three problems: a very heavy computational cost due to dense input, a disputable pooling operation in modeling relations of the visual cortex, and blind feature selection in a feed-forward framework. To solve these problems, we develop an enhanced BIM (EBIM), which removes uninformative input by imposing sparsity constraints, utilizes a novel local weighted pooling operation with stronger physiological motivations, and applies a feedback procedure that selects effective features for combination. Empirical studies on the CalTech5 database and CalTech101 database show that EBIM is more effective and efficient than BIM. We also apply EBIM to the MIT-CBCL street scene database to show it achieves comparable performance in comparison with the current best performance. Moreover, the new system can process images with resolution 128 times 128 at a rate of 50 frames per second and enhances the speed 20 times at least in comparison with BIM in common applications.}, author = {Huang, Yongzhen and Huang, Kaiqi and Wang, Liangsheng and Tao, Dacheng and Tan, Tieniu and Li, Xuelong}, citeulike-article-id = {5816896}, citeulike-linkout-0 = {http://dx.doi.org/10.1109/CVPR.2008.4587599}, citeulike-linkout-1 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4587599}, day = {05}, doi = {10.1109/CVPR.2008.4587599}, journal = {Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on}, keywords = {mitchell-group, object-recognition}, month = {August}, pages = {1--8}, posted-at = {2009-09-21 23:46:27}, priority = {2}, title = {Enhanced biologically inspired model}, url = {http://dx.doi.org/10.1109/CVPR.2008.4587599}, year = {2008} }

TY - CONF ID - citeulike:5816896 L3 - citeulike-article-id:5816896 N2 - It has been demonstrated by Serre et al. that the biologically inspired model (BIM) is effective for object recognition. It outperforms many state-of-the-art methods in challenging databases. However, BIM has the following three problems: a very heavy computational cost due to dense input, a disputable pooling operation in modeling relations of the visual cortex, and blind feature selection in a feed-forward framework. To solve these problems, we develop an enhanced BIM (EBIM), which removes uninformative input by imposing sparsity constraints, utilizes a novel local weighted pooling operation with stronger physiological motivations, and applies a feedback procedure that selects effective features for combination. Empirical studies on the CalTech5 database and CalTech101 database show that EBIM is more effective and efficient than BIM. We also apply EBIM to the MIT-CBCL street scene database to show it achieves comparable performance in comparison with the current best performance. Moreover, the new system can process images with resolution 128 times 128 at a rate of 50 frames per second and enhances the speed 20 times at least in comparison with BIM in common applications. JF - Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on EP - 8 TI - Enhanced biologically inspired model SP - 1 KW - mitchell-group KW - object-recognition AU - Huang, Yongzhen AU - Huang, Kaiqi AU - Wang, Liangsheng AU - Tao, Dacheng AU - Tan, Tieniu AU - Li, Xuelong PY - 2008/08/05/ UR - http://dx.doi.org/10.1109/CVPR.2008.4587599 ER -