Fast multi-image matching via density-based clustering
MetadataShow full item record
Accepted manuscriptSupporting documentation
Citation (published version)R. Tron, X. Zhou, C. Esteves, K. Daniilidis. 2017. "Fast Multi-Image Matching via Density-Based Clustering." Proceedings / IEEE International Conference on Computer Vision. IEEE International Conference on Computer Vision. IEEE International Conference on Computer Vision. https://doi.org/10.1109/ICCV.2017.437
We consider the problem of finding consistent matches across multiple images. Previous state-of-the-art solutions use constraints on cycles of matches together with convex optimization, leading to computationally intensive iterative algorithms. In this paper, we propose a clustering-based formulation. We first rigorously show its equivalence with the previous one, and then propose QuickMatch, a novel algorithm that identifies multi-image matches from a density function in feature space. We use the density to order the points in a tree, and then extract the matches by breaking this tree using feature distances and measures of distinctiveness. Our algorithm outperforms previous state-of-the-art methods (such as MatchALS) in accuracy, and it is significantly faster (up to 62 times faster on some bechmarks), and can scale to large datasets (with more than twenty thousands features).