Research map / Detection and segmentation / CNN features
CircleNet for Hip Landmark Detection
Hai Wu, Hongtao Xie, Chuanbin Liu, Zheng-Jun Zha, Jun Sun, Yongdong Zhang
Read the paperCheck 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 |
|---|---|---|---|
| add LNL block after stage 4 (ablation) | CircleNet · DDH test | average error (pixels): 4.99 -> 4.79 | Improved |
| add LNL block after stage 4 (ablation) | CircleNet · DDH test | FPS: 25.6 -> 22.7 | Hurt |
| use CircleNet radius prediction vs Faster R-CNN+GN+WS | Faster R-CNN · DDH test | average error (pixels): 6.59 -> 4.99 | Improved |
| place LNL after stage 5 instead of stage 4 (ablation) | CircleNet+LNL · DDH test | average error (pixels): 4.79 -> 4.96 | Hurt |
| place LNL after stage 5 instead of stage 4 (ablation) | CircleNet+LNL · DDH test | FPS: 22.7 -> 11.8 | Hurt |
| add non-local (NL) block after stage 4 (ablation) | CircleNet · DDH test | average error (pixels): 4.99 -> 4.95 | No clear effect |
| add GC block after stage 4 (ablation) | CircleNet · DDH test | average error (pixels): 4.99 -> 4.86 | No clear effect |
| radius loss weight lambda_r 0.01 to 0.1 (ablation) | CircleNet+LNL · DDH test | average error (pixels): 4.90 -> 4.79 | Improved |
| vary lambda_r among 0.2, 0.5, 0.7, 1 (ablation) | CircleNet+LNL · DDH test | average error (pixels): 4.79 vs 4.92, 4.82, 4.83, 4.90 | Hurt |
| area ratio mu 20/32 to 24/32 (ablation) | LNL block · DDH test | average error (pixels): 4.84 -> 4.79 | Improved |
| area ratio mu 28/32 instead of 24/32 (ablation) | LNL block · DDH test | average error (pixels): 4.79 -> 4.92 | Hurt |
| Gaussian instead of Embedded Gaussian relation (ablation) | LNL block · DDH test | average error (pixels): 4.79 -> 4.99 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- Libra R-CNN — Libra R-CNN: Towards Balanced Learning for Object Detection (CVPR 2019) · better on 1 of 1 shared result
- FCOS — FCOS: Fully Convolutional One-Stage Object Detection (ICCV 2019) · better on 1 of 1 shared result
- Cascade R-CNN — Cascade R-CNN: Delving Into High Quality Object Detection (CVPR 2018) · better on 1 of 1 shared result
- Faster R-CNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · better on 1 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
- Contour-Hugging Heatmaps for Landmark Detection (CVPR 2022 · similarity 0.84)
- Window Loss for Bone Fracture Detection and Localization in X-ray Images with Point-based Annotation (AAAI 2021 · similarity 0.84)
- Domain Generalized Medical Landmark Detection via Robust Boundary-Aware Pre-Training (AAAI 2025 · similarity 0.83)
- DPA-P2PNet: Deformable Proposal-Aware P2PNet for Accurate Point-Based Cell Detection (AAAI 2024 · similarity 0.81)
- VISTA3D: A Unified Segmentation Foundation Model For 3D Medical Imaging (CVPR 2025 · similarity 0.80)
- T-FAKE: Synthesizing Thermal Images for Facial Landmarking (CVPR 2025 · similarity 0.80)
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.
