D2D: Keypoint Extraction with Describe to Detect Approach
arXiv:2005.13605
Abstract
In this paper, we present a novel approach that exploits the information within the descriptor space to propose keypoint locations. Detect then describe, or detect and describe jointly are two typical strategies for extracting local descriptors. In contrast, we propose an approach that inverts this process by first describing and then detecting the keypoint locations. % Describe-to-Detect (D2D) leverages successful descriptor models without the need for any additional training. Our method selects keypoints as salient locations with high information content which is defined by the descriptors rather than some independent operators. We perform experiments on multiple benchmarks including image matching, camera localisation, and 3D reconstruction. The results indicate that our method improves the matching performance of various descriptors and that it generalises across methods and tasks.
References in corpus (10)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- R2D2: Repeatable and Reliable Detector and Descriptor
- TILDE: A Temporally Invariant Learned DEtector
- GeoDesc: Learning Local Descriptors by Integrating Geometry Constraints
- UnsuperPoint: End-to-end Unsupervised Interest Point Detector and Descriptor
- Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters
- ASLFeat: Learning Local Features of Accurate Shape and Localization
- Large scale evaluation of local image feature detectors on homography datasets
- A Large Dataset for Improving Patch Matching
- S2DNet: Learning Accurate Correspondences for Sparse-to-Dense Feature Matching