Class segmentation and object localization with superpixel neighborhoods

Brian Fulkerson, Andrea Vedaldi, Stefano Soatto
2009 2009 IEEE 12th International Conference on Computer Vision  
We propose a method to identify and localize object classes in images. Instead of operating at the pixel level, we advocate the use of superpixels as the basic unit of a class segmentation or pixel localization scheme. To this end, we construct a classifier on the histogram of local features found in each superpixel. We regularize this classifier by aggregating histograms in the neighborhood of each superpixel and then refine our results further by using the classifier in a conditional random
more » ... eld operating on the superpixel graph. Our proposed method exceeds the previously published state-of-the-art on two challenging datasets: Graz-02 and the PASCAL VOC 2007 Segmentation Challenge.
doi:10.1109/iccv.2009.5459175 dblp:conf/iccv/FulkersonVS09 fatcat:wcgp3w3bvvaofnnj62uq3ytb34