Title: SOLAR: Second-Order Loss and Attention for Image Retrieval

Abstract: Recent works in deep-learning have shown that utilising second-order information is beneficial in many computer-vision related tasks. Second-order information can be enforced both in the spatial context and the abstract feature dimensions. In this work we explore two second order components. One is focused on second-order spatial information to increase the performance of image descriptors, both local and global. More specifically, it is used to re-weight feature maps, and thus emphasise salient image locations that are subsequently used for description. The second component is concerned with a second-order similarity (SOS) loss, that we extend to global descriptors for image retrieval, and is used to enhance the triplet loss with hard negative mining. We validate our approach on two different tasks and three datasets for image retrieval and patch matching. The results show that our second order components bring significant performance improvements in both tasks and lead to state of the art results across the benchmarks.