Optimal Geometric Matching for Patch-Based Object Detection

Daniel Keysers, Thomas Deselaers, Thomas Breuel
Electronic Letters on Computer Vision and Image Analysis volume 6 number 1, Pages 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.




@article{ KEYS2007,
	Title = {Optimal Geometric Matching for Patch-Based Object Detection},
	Author = {Daniel Keysers and Thomas Deselaers and Thomas Breuel},
	Year = {2007},
	Publisher = {CVC Press},
	Publisher = {6},
	Pages = {44-54},
	Journal = {Electronic Letters on Computer Vision and Image Analysis},
	Number = {1}

Last modified:: 30.08.2016