The Community for Technology Leaders
2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017)
Honolulu, Hawaii, USA
July 21, 2017 to July 26, 2017
ISSN: 1063-6919
ISBN: 978-1-5386-0457-1
pp: 3929-3937
ABSTRACT
Several machine learning tasks require to represent the data using only a sparse set of interest points. An ideal detector is able to find the corresponding interest points even if the data undergo a transformation typical for a given domain. Since the task is of high practical interest in computer vision, many hand-crafted solutions were proposed. In this paper, we ask a fundamental question: can we learn such detectors from scratch? Since it is often unclear what points are interesting, human labelling cannot be used to find a truly unbiased solution. Therefore, the task requires an unsupervised formulation. We are the first to propose such a formulation: training a neural network to rank points in a transformation-invariant manner. Interest points are then extracted from the top/bottom quantiles of this ranking. We validate our approach on two tasks: standard RGB image interest point detection and challenging cross-modal interest point detection between RGB and depth images. We quantitatively show that our unsupervised method performs better or on-par with baselines.
INDEX TERMS
image colour analysis, neural nets, object detection, unsupervised learning
CITATION

N. Savinov, A. Seki, L. Ladicky, T. Sattler and M. Pollefeys, "Quad-Networks: Unsupervised Learning to Rank for Interest Point Detection," 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, Hawaii, USA, 2017, pp. 3929-3937.
doi:10.1109/CVPR.2017.418
84 ms
(Ver 3.3 (11022016))