D2-Net swMATH ID: 42672 Software Authors: Mihai Dusmanu, Ignacio Rocco, Tomas Pajdla, Marc Pollefeys, Josef Sivic, Akihiko Torii, Torsten Sattler Description: D2-Net: A Trainable CNN for Joint Detection and Description of Local Features. In this work we address the problem of finding reliable pixel-level correspondences under difficult imaging conditions. We propose an approach where a single convolutional neural network plays a dual role: It is simultaneously a dense feature descriptor and a feature detector. By postponing the detection to a later stage, the obtained keypoints are more stable than their traditional counterparts based on early detection of low-level structures. We show that this model can be trained using pixel correspondences extracted from readily available large-scale SfM reconstructions, without any further annotations. The proposed method obtains state-of-the-art performance on both the difficult Aachen Day-Night localization dataset and the InLoc indoor localization benchmark, as well as competitive performance on other benchmarks for image matching and 3D reconstruction. Homepage: https://arxiv.org/abs/1905.03561 Source Code: https://github.com/mihaidusmanu/d2-net Related Software: SuperPoint; HPatches; BRISK; R2D2; SuperGlue; ContextDesc; ASLFeat; LF-Net; LoFTR; Patch2Pix; PointNet; SOSNet; L2-Net; Theia; InLoc; ORB-SLAM2; ORB-SLAM; MAGSAC++; COTR; MS-COCO Cited in: 3 Publications all top 5 Cited by 10 Authors 1 Fan, Aoxiang 1 Gong, Mingming 1 Jiang, Junjun 1 Jiang, Xingyu 1 Ma, Jiayi 1 Szeliski, Richard 1 Tao, Dacheng 1 Yan, Junchi 1 You, Xinge 1 Zheng, Qi Cited in 2 Serials 2 International Journal of Computer Vision 1 Texts in Computer Science Cited in 3 Fields 3 Computer science (68-XX) 1 Numerical analysis (65-XX) 1 Information and communication theory, circuits (94-XX) Citations by Year