PASCAL VOC segmentation challenge 2009

Team: Xavier Boix, Josep Gonfaus, Fahad Shahbaz Khan, Joost van de Weijer, Andrew Bagdanov, Marco Pedersoli, Jordi González, Joan Serrat.

Method: We use a conditional random field (CRF) approach. The main novelty is the introduction of a new potential, called harmony potential, which allows to encode any combinations of labels at a node. This improvement is especially relevant for the larger scale regions in the image, and allows us to exploit to results of our image classification method. The method is explained in detail in our CVPR 2010 paper.

Results: Our method obtained the best score in 6 of the 20 classes, thereby finishing second behind the Bonn submission. More results can be found on the VOC workshop page.

Literature

Josep Gonfaus, Xavier Boix, Joost van de Weijer, Anrew Bagdanov, Joan Serrat, Jordi González, Harmony Potentials for Joint Classification and Segmentation, Proc. CVPR 2010, San Fransisico, USA.