Learning Affinity-Aware Upsampling for Deep Image Matting

We show that learning affinity in upsampling provides an effective and efficient approach to exploit pairwise interactions in deep networks. Second-order features are commonly used in dense prediction to build adjacent relations with a learnable module after upsampling such as non-local blocks. Since upsampling is essential, learning affinity in upsampling can avoid additional propagation layers, offering the potential for building compact models. By looking at existing upsampling operators from a unified mathematical perspective, we generalize them into a second-order form and introduce Affinity-Aware Upsampling (A(2)U) where upsampling kernels are generated using a light-weight low-rank bilinear model and are conditioned on second-order features. Our upsampling operator can also be extended to downsampling. We discuss alternative implementations of A(2)U and verify their effectiveness on two detail-sensitive tasks: image reconstruction on a toy dataset; and a large-scale image matting task where affinity-based ideas constitute mainstream matting approaches. In particular, results on the Composition-1k matting dataset show that A(2)U achieves a 14% relative improvement in the SAD metric against a strong baseline with negligible increase of parameters ( < 0.5%). Compared with the state-of-the-art matting network, we achieve 8% higher performance with only 40% model complexity.