Probabilistic Warp Consistency for Weakly-Supervised Semantic Correspondences
This addresses the challenge of semantic correspondence estimation in computer vision, particularly for object matching across images, with incremental improvements over existing methods.
The paper tackles the problem of weakly-supervised semantic matching by proposing Probabilistic Warp Consistency, which supervises dense matching scores as a conditional probability distribution and handles occlusions with a learnable unmatched state. The method sets a new state-of-the-art on four benchmarks and improves performance in strongly-supervised settings when combined with keypoint annotations.
We propose Probabilistic Warp Consistency, a weakly-supervised learning objective for semantic matching. Our approach directly supervises the dense matching scores predicted by the network, encoded as a conditional probability distribution. We first construct an image triplet by applying a known warp to one of the images in a pair depicting different instances of the same object class. Our probabilistic learning objectives are then derived using the constraints arising from the resulting image triplet. We further account for occlusion and background clutter present in real image pairs by extending our probabilistic output space with a learnable unmatched state. To supervise it, we design an objective between image pairs depicting different object classes. We validate our method by applying it to four recent semantic matching architectures. Our weakly-supervised approach sets a new state-of-the-art on four challenging semantic matching benchmarks. Lastly, we demonstrate that our objective also brings substantial improvements in the strongly-supervised regime, when combined with keypoint annotations.