Research map / Label-efficient and robust learning / Test-time adaptation
Topology-Inspired Backward-Free Framework for Test-Time Adaptation in Medical Detection
Bin Pu, Xingguo Lv, Jiewen Yang, Kai Xu, Lei Zhao, Zuozhu Liu, Kenli Li
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 test-time augmentation (flip, rotate, noise, crop, resize) with NMS fusion (ablation) | T3A · FUSH2 heart Center 1->2 | mAP: 62.93 -> 72.46 | Improved |
| add Gaussian Kernel Fusion (GKF) before NMS (ablation) | T3A · FUSH2 heart Center 1->2 | mAP: 72.46 -> 73.20 | Improved |
| add GKF before NMS (ablation) | T3A · FUSH2 heart Center 2->1 | mAP: 86.25 -> 86.09 | Hurt |
| replace NMS with weighted boxes fusion (ablation) | T3A · FUSH2 heart Center 1->2 | mAP: 73.20 -> 73.44 | Improved |
| replace NMS with BRA (heatmap-guided box fusion) (ablation) | T3A · FUSH2 heart Center 1->2 | mAP: 73.20 -> 75.57 | Improved |
| replace NMS with BRA (ablation) | T3A · FUSH2 heart Center 2->1 | mAP: 86.09 -> 89.37 | Improved |
| add flip, rotate, noise, crop and more resize scales (ablation) | T3A · FUSH2 heart Center 1->2 | mAP: 62.93 -> 72.46 | Improved |
| vary Gaussian kernel scale alpha (0.1, 0.2, 0.5, 1.0) (ablation) | T3A · FUSH2 heart Center 1->2 | mAP: 75.08, 75.57, 75.29, 74.61 | No clear effect |
| use T3A instead of backprop-based TTA (TENT) | source detector · FUSH2 heart Center 1->2 | mAP: 46.70 -> 75.57 | Improved |
| apply backprop-based TTA methods (TENT, DLTTA, MonoTTA) | source detector · FCS heart Center B->A | mAP: No Adapt 77.16; TENT 80.28; MonoTTA 34.08 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TENT — Tent: Fully Test-Time Adaptation by Entropy Minimization (ICLR 2021) · better on 2 of 2 shared results
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
- ToMo-UDA++: Unsupervised Domain Adaptation for Anatomical Structure Detection Using Enhanced Topology and Morphology Knowledge (IJCV 2026 · similarity 0.87)
- Unsupervised Domain Adaptation for Anatomical Structure Detection in Ultrasound Images (ICML 2024 · similarity 0.87)
- Leveraging Anatomical Consistency for Multi-Object Detection in Ultrasound Images via Source-free Unsupervised Domain Adaptation (AAAI 2025 · similarity 0.86)
- Anatomical Domain Shifts: Test-time Heterogeneous Adaptation for 3D Human Pose Prediction (CVPR 2026 · similarity 0.86)
- Gradient Alignment Improves Test-Time Adaptation for Medical Image Segmentation (AAAI 2025 · similarity 0.84)
- Test-Time Adaptation via Style and Structure Guidance for Histological Image Registration (AAAI 2024 · similarity 0.84)
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.
