Binary coherent edge descriptors

Patch descriptors are used for a variety of tasks ranging from finding corresponding points across images, to describing object category parts. In this paper, we propose an image patch descriptor based on edge position, orientation and local linear length. Unlike previous works using histograms of gradients, our descriptor does not encode relative gradient magnitudes. Our approach locally normalizes the patch gradients to remove relative gradient information, followed by orientation dependent binning. Finally, the edge histogram is binarized to encode edge locations, orientations and lengths. Two additional extensions are proposed for fast PCA dimensionality reduction, and a min-hash approach for fast patch retrieval. Our algorithm produces state-of-the-art results on previously published object instance patch data sets, as well as a new patch data set modeling intra-category appearance variations.

• Type: Conference
• Year: 2010
• Where: ECCV
• Keywords: Computer Vision, ECCV 2010, Image Patch Descriptor, Patch Data, Patch Descriptor