@conference {Zheng2014, title = {Dense semantic image segmentation with objects and attributes}, booktitle = {Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition}, year = {2014}, pages = {3214{\textendash}3221}, abstract = {The concepts of objects and attributes are both important for describing images precisely, since verbal descriptions often contain both adjectives and nouns (e.g. {\textquoteright}I see a shiny red chair{\textquoteright}). In this paper, we formulate the problem of joint visual attribute and object class image segmentation as a dense multi-labelling problem, where each pixel in an image can be associated with both an object-class and a set of visual attributes labels. In order to learn the label correlations, we adopt a boosting-based piecewise training approach with respect to the visual appearance and co-occurrence cues. We use a filtering-based mean-field approximation approach for efficient joint inference. Further, we develop a hierarchical model to incorporate region-level object and attribute information. Experiments on the aPASCAL, CORE and attribute augmented NYU indoor scenes datasets show that the proposed approach is able to achieve state-of-the-art results.}, keywords = {Attributes, Image segmentation, Object recognition, Scene Understanding}, isbn = {9781479951178}, issn = {10636919}, doi = {10.1109/CVPR.2014.411}, url = {http://www.robots.ox.ac.uk/\~{}tvg/http://tu-dresden.de/inf/cvld}, author = {Zheng, Shuai and Cheng, Ming Ming and Warrell, Jonathan and Sturgess, Paul and Vineet, Vibhav and Carsten Rother and Torr, Philip H.S.} }