Kernel-Predictability: A New Information Measure and Its Application to Image Registration
Abstract
A new information measure for probability distributions is presented; based on it, a similarity measure between images is derived, which is used for constructing a robust image registration algorithm based on random sampling, similar to classical approaches like mutual information. It is shown that the registration method obtained with the new similarity measure shows a significantly better performance for small sampling sets; this makes it specially suited for the estimation of non-parametric deformation fields, where the estimation of the local transformation is done on small windows. This is confirmed by extensive comparisons using synthetic deformations of real images.
Cite
Text
García et al. "Kernel-Predictability: A New Information Measure and Its Application to Image Registration." European Conference on Computer Vision, 2006. doi:10.1007/11744078_39Markdown
[García et al. "Kernel-Predictability: A New Information Measure and Its Application to Image Registration." European Conference on Computer Vision, 2006.](https://mlanthology.org/eccv/2006/garcia2006eccv-kernel/) doi:10.1007/11744078_39BibTeX
@inproceedings{garcia2006eccv-kernel,
title = {{Kernel-Predictability: A New Information Measure and Its Application to Image Registration}},
author = {García, Héctor Fernando Gómez and Marroquín, José L. and Van Horebeek, Johan},
booktitle = {European Conference on Computer Vision},
year = {2006},
pages = {502-513},
doi = {10.1007/11744078_39},
url = {https://mlanthology.org/eccv/2006/garcia2006eccv-kernel/}
}