CPPP/UFMS at ImageCLEF 2014: Robot Vision Task

Abstract. This paper describes the participation of the CPPP/UFMS group in the robot vision task. We have applied the spatial pyramid matching proposed by Lazebnik et al. This method extends bag-of-visual-words to spatial pyramids by concatenating histograms of local features found in increasingly fine sub-regions. To form the visual vocabulary, k-means clustering was applied in a random subset of images from training dataset. After that the images are classified using a pyramid match kernel and the k-nearest neighbors. The system has shown promising results, particularly for object recognition. Key words: Scene recognition, object recognition, spatial pyramid match-ing