Affine Subspace Representation for Feature Description

TL;DR

The paper introduces the Affine Subspace Representation (ASR) descriptor, which robustly encodes local features under affine distortions caused by viewpoint changes, outperforming traditional descriptors like SIFT.

Contribution

It presents a novel subspace-based descriptor that captures multi-view local information efficiently and accurately, with an accelerated computation method for practical use.

Findings

ASR outperforms state-of-the-art descriptors under various image transformations.

ASR performs well without a dedicated affine invariant detector.

The method effectively encodes local structures across multiple views.

Abstract

This paper proposes a novel Affine Subspace Representation (ASR) descriptor to deal with affine distortions induced by viewpoint changes. Unlike the traditional local descriptors such as SIFT, ASR inherently encodes local information of multi-view patches, making it robust to affine distortions while maintaining a high discriminative ability. To this end, PCA is used to represent affine-warped patches as PCA-patch vectors for its compactness and efficiency. Then according to the subspace assumption, which implies that the PCA-patch vectors of various affine-warped patches of the same keypoint can be represented by a low-dimensional linear subspace, the ASR descriptor is obtained by using a simple subspace-to-point mapping. Such a linear subspace representation could accurately capture the underlying information of a keypoint (local structure) under multiple views without sacrificing its distinctiveness. To accelerate the computation of ASR descriptor, a fast approximate algorithm is proposed by moving the most computational part (ie, warp patch under various affine transformations) to an offline training stage. Experimental results show that ASR is not only better than the state-of-the-art descriptors under various image transformations, but also performs well without a dedicated affine invariant detector when dealing with viewpoint changes.