%0 Conference Proceedings
%T Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters
%A Axel Barroso-Laguna
%A Edgar Riba
%A Daniel Ponsa
%A Krystian Mikolajczyk
%B 18th IEEE International Conference on Computer Vision
%D 2019
%F Axel Barroso-Laguna2019
%O MSIAU; 600.122
%O exported from refbase (http://refbase.cvc.uab.es/show.php?record=3290), last updated on Mon, 07 Dec 2020 14:18:41 +0100
%X We introduce a novel approach for keypoint detection task that combines handcrafted and learned CNN filters within a shallow multi-scale architecture. Handcrafted filters provide anchor structures for learned filters, which localize, score and rank repeatable features. Scale-space representation is used within the network to extract keypoints at different levels. We design a loss function to detect robust features that exist across a range of scales and to maximize the repeatability score. Our Key.Net model is trained on data synthetically created from ImageNet and evaluated on HPatches benchmark. Results show that our approach outperforms state-of-the-art detectors in terms of repeatability, matching performance and complexity.
%U https://ieeexplore.ieee.org/document/9010664
%U http://refbase.cvc.uab.es/files/BRP2019.pdf
%U http://dx.doi.org/10.1109/ICCV.2019.00593
%P 5835-5843