Affine Subspace Representation for Feature Description

European Conference on Computer Vision (ECCV 2014)

Zhenhua Wang, Bin Fan and Fuchao Wu





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 infor- mation of multi-view patches, making it robust to affine distortions while main- taining 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 sim- ple subspace-to-point mapping. Such a linear subspace representation could ac- curately capture the underlying information of a keypoint (local structure) under multiple views without sacrificing its distinctiveness. To accelerate the computa- tion of ASR descriptor, a fast approximate algorithm is proposed by moving the most computational part (i.e., 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 al- so performs well without a dedicated affine invariant detector when dealing with viewpoint changes.


Some Results


For more detailed results, please refer to the paper


Poster Spotlight:


Paper: pdf
BitTex: bib


If you have any questions about this paper, please contact the author (wzh[at]