Kernel Local Descriptors with Implicit Rotation Matching
Abstract
In this work we design a kernelized local feature descriptor and propose a matching scheme for aligning patches quickly and automatically. We analyze the SIFT descriptor from a kernel view and identify and reproduce some of its underlying benefits. We overcome the quantization artifacts of SIFT by encoding pixel attributes in a continuous manner via explicit feature maps. Experiments performed on the patch dataset of Brown et al. [3] show the superiority of our descriptor over methods based on supervised learning.
Fichier principal
BursucToliasJegou_ICMR2015_Kernel Local Descriptors with Implicit Rotation Matching.pdf (440.89 Ko)
Télécharger le fichier
Origin | Files produced by the author(s) |
---|
Loading...