LOCUS: learning object classes with unsupervised segmentation

@inproceedings{Winn:ICCV05, abstract = {We address the problem of learning object class models and object segmentations from unannotated images. We introduce LOCUS (learning object classes with unsupervised segmentation) which uses a generative probabilistic model to combine bottom-up cues of color and edge with top-down cues of shape and pose. A key aspect of this model is that the object appearance is allowed to vary from image to image, allowing for significant within-class variation. By iteratively updating the belief in the object's position, size, segmentation and pose, LOCUS avoids making hard decisions about any of these quantities and so allows for each to be refined at any stage. We show that LOCUS successfully learns an object class model from unlabeled images, whilst also giving segmentation accuracies that rival existing supervised methods. Finally, we demonstrate simultaneous recognition and segmentation in novel images using the learned models for a number of object classes, as well as unsupervised object discovery and tracking in video.}, author = {Winn, J. and Jojic, N. }, citeulike-article-id = {1139902}, doi = {http://dx.doi.org/10.1109/ICCV.2005.148}, journal = {Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on}, keywords = {nips2008}, pages = {756--763 Vol. 1}, posted-at = {2008-06-06 17:26:00}, priority = {2}, title = {LOCUS: learning object classes with unsupervised segmentation}, url = {http://dx.doi.org/10.1109/ICCV.2005.148}, volume = {1}, year = {2005} }

TY - CONF ID - Winn:ICCV05 L3 - citeulike-article-id:1139902 TI - LOCUS: learning object classes with unsupervised segmentation JF - Computer Vision, 2005. ICCV 2005. Tenth IEEE International Conference on VL - 1 SP - 756 EP - 763 Vol. 1 N2 - We address the problem of learning object class models and object segmentations from unannotated images. We introduce LOCUS (learning object classes with unsupervised segmentation) which uses a generative probabilistic model to combine bottom-up cues of color and edge with top-down cues of shape and pose. A key aspect of this model is that the object appearance is allowed to vary from image to image, allowing for significant within-class variation. By iteratively updating the belief in the object's position, size, segmentation and pose, LOCUS avoids making hard decisions about any of these quantities and so allows for each to be refined at any stage. We show that LOCUS successfully learns an object class model from unlabeled images, whilst also giving segmentation accuracies that rival existing supervised methods. Finally, we demonstrate simultaneous recognition and segmentation in novel images using the learned models for a number of object classes, as well as unsupervised object discovery and tracking in video. KW - nips2008 AU - Winn, J AU - Jojic, N PY - 2005/// UR - http://dx.doi.org/10.1109/ICCV.2005.148 ER -