KLUCKNER et al.: CONSISTENT REGIONS AND INDIVIDUAL CONTEXT 1 Semantic Image Classification Using Consistent Regions and Individual Context

This paper proposes an efficient approach for semantic image classification by inte-grating additional contextual constraints such as class co-occurrences into a randomized forest classification framework. The randomized forest classifier performs an initial yet local classification on the pixel level by using powerful covariance matrix based de-scriptors as feature representation. Furthermore, we exploit multiple unsupervised im-age partitions to provide a reliable spatial region support and to capture the real object boundaries. An information theoretic driven approach detects consistently classified re-gions and generates a representative segmentation incorporating the classification result on the pixel level. Moreover, we use a conditional random field formulation to obtain a final labeling including context information individually generated for each test image. To illustrate state-of-the-art performance, we run experiments on the two versions of the MSRC [21] dataset with 9 and 21 object classes and on the PASCAL VOC2007 [5] image collection.