Abstract

We address the problem of semantic segmentation, or multi- class pixel labeling, by constructing a graph of dense overlapping patch correspondences across large image sets. We then transfer annotations from labeled images to unlabeled images using the established patch correspondences. Unlike previous approaches to non-parametric label transfer our approach does not require an initial image retrieval step. Moreover, we operate on a graph for computing mappings between im- ages, which avoids the need for exhaustive pairwise comparisons. Con- sequently, we can leverage offline computation to enhance performance at test time. We conduct extensive experiments to analyze different vari- ants of our graph construction algorithm and evaluate multi-class pixel labeling performance on several challenging datasets.