Affinity CNN: Learning Pixel-Centric Pairwise Relations for Figure/Ground Embedding.

Michael Maire, Takuya Narihira, Stella X. Yu

Browse the full CVPR paper archive.