Multi-View Feature Engineering and Learning

Jingming Dong, Nikolaos Karianakis, Damek Davis, Joshua Hernandez, Jonathan Balzer, Stefano Soatto; The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 3251-3260


We frame the problem of local representation of imaging data as the computation of minimal sufficient statistics that are invariant to nuisance variability induced by viewpoint and illumination. We show that, under very stringent conditions, these are related to "feature descriptors" commonly used in Computer Vision. Such conditions can be relaxed if multiple views of the same scene are available. We propose a sampling-based and a point-estimate based approximation of such a representation, compared empirically on image-to-(multiple)image matching, for which we introduce a multi-view wide-baseline matching benchmark, consisting of a mixture of real and synthetic objects with ground truth camera motion and dense three-dimensional geometry.

Related Material

author = {Dong, Jingming and Karianakis, Nikolaos and Davis, Damek and Hernandez, Joshua and Balzer, Jonathan and Soatto, Stefano},
title = {Multi-View Feature Engineering and Learning},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2015}