Affine steerers for structured keypoint description
This work addresses the need for robust image matching in computer vision by improving descriptor invariance to affine distortions, though it is incremental as it builds on existing steerer concepts.
The paper tackled the problem of making deep learning-based keypoint descriptors approximately equivariant to locally affine image transformations by generalizing steerers from rotations to affine transformations using GL(2) representation theory, resulting in state-of-the-art performance on standard benchmarks.
We propose a way to train deep learning based keypoint descriptors that makes them approximately equivariant for locally affine transformations of the image plane. The main idea is to use the representation theory of GL(2) to generalize the recently introduced concept of steerers from rotations to affine transformations. Affine steerers give high control over how keypoint descriptions transform under image transformations. We demonstrate the potential of using this control for image matching. Finally, we propose a way to finetune keypoint descriptors with a set of steerers on upright images and obtain state-of-the-art results on several standard benchmarks. Code will be published at github.com/georg-bn/affine-steerers.