Research map / Detection and segmentation / Weakly supervised detection and segmentation
Decoupling Makes Weakly Supervised Local Feature Better
Kunhong Li, Longguang Wang, Li Liu, Qing Ran, Kai Xu, Yulan Guo
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| decouple descriptor and detector training, learning descriptors before training the detector (ablation) | weakly supervised local feature learning · HPatches | matching accuracy | Improved |
| jointly optimize the description and detection networks (ablation) | weakly supervised local feature learning · HPatches viewpoint changes | matching accuracy | Hurt |
| use SIFT keypoints instead of learned keypoints (ablation) | PoSFeat feature description · HPatches | matching accuracy | Mixed |
| replace line-to-window search with coarse-to-fine search (ablation) | descriptor learning with SIFT keypoints · HPatches | matching accuracy | Hurt |
| increase the number of epipolar-line samples or use a larger local window (ablation) | PoSFeat descriptor learning · HPatches | MMAscore | Improved |
| use PoSFeat instead of DISK-W | weakly supervised local features · HPatches illumination changes | MMAscore: 0.803 -> 0.826 | Improved |
| use PoSFeat instead of DISK-W | weakly supervised local features · HPatches viewpoint changes | MMAscore: 0.649 -> 0.728 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- R2D2 — R2D2: Reliable and Repeatable Detector and Descriptor (NeurIPS 2019) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Learning Feature Descriptors using Camera Pose Supervision (ECCV 2020 · similarity 0.87)
- W2F: A Weakly-Supervised to Fully-Supervised Framework for Object Detection (CVPR 2018 · similarity 0.87)
- Object-Aware Instance Labeling for Weakly Supervised Object Detection (ICCV 2019 · similarity 0.86)
- Salvage of Supervision in Weakly Supervised Object Detection and Segmentation (TPAMI 2023 · similarity 0.86)
- Weakly Supervised Object Localization and Detection: A Survey (TPAMI 2021 · similarity 0.86)
- Cap2Det: Learning to Amplify Weak Caption Supervision for Object Detection (ICCV 2019 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
