Optimal Geometric Matching for Patch-Based Object Detection

Daniel Keysers, Thomas Deselaers, Thomas Breuel

In: Electronic Letters on Computer Vision and Image Analysis (ELCVIA) 6 1 Seiten 44-54 CVC Press 2007.


We present an efficient method to determine the optimal matching of two patch-based image object representations under rotation, scaling, and translation (RST). This use of patches is equivalent to a fully-connected part-based model, for which the presented approach offers an efficient procedure to determine the best fit. While other approaches that use fully connected models have a high complexity in the number of parts used, we achieve linear complexity in that variable, because we only allow RST-matchings. The presented approach is used for object recognition in images: by matching images that contain certain objects to a test image, we can detect whether the test image contains an object of that class or not. We evaluate this approach on the Caltech data and obtain very competitive results.

136-208-1-PB.pdf (pdf, 732 KB )

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence