Research map / Detection and segmentation / CNN features
Anatomy-Aware Siamese Network: Exploiting Semantic Asymmetry for Accurate Pelvic Fracture Detection in X-ray Images
Haomin Chen, Yirui Wang, Kang Zheng, Weijian Li, Chi-Tung Chang, Adam P. Harrison, Jing Xiao, Gregory D. Hager, Le Lu, Chien-Hung Liao, Shun Miao
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 symmetric feature fusion (ablation) | vanilla DenseNet-121 · Pelvic X-ray fracture detection (5-fold CV) | Recall at FP=1%: 70.79% -> 73.22% | Improved |
| add feature alignment on top of feature fusion (ablation) | AASN FF · Pelvic X-ray fracture detection (5-fold CV) | Recall at FP=1%: 73.22% -> 76.68% | Improved |
| add feature alignment on top of FF+CL (ablation) | AASN FF+CL · Pelvic X-ray fracture detection (5-fold CV) | Recall at FP=1%: 76.27% -> 77.87% | Improved |
| add contrastive learning on top of FF (ablation) | AASN FF · Pelvic X-ray fracture detection (5-fold CV) | Recall at FP=1%: 73.22% -> 76.27% | Improved |
| add contrastive learning on top of FF+FA (ablation) | AASN FF+FA · Pelvic X-ray fracture detection (5-fold CV) | Recall at FP=1%: 76.68% -> 77.87% | Improved |
| add contrastive learning on top of FF+FA (ablation) | AASN FF+FA · Pelvic X-ray fracture detection (5-fold CV) | AP: 95.68% -> 96.50% | Improved |
| add contrastive learning on top of FF+FA (ablation) | AASN FF+FA · Pelvic X-ray fracture detection (5-fold CV) | AUC: 97.20% -> 97.71% | Improved |
| add non-linear projection head to contrastive learning (ablation) | AASN · Pelvic X-ray fracture detection (5-fold CV) | Recall at FP=1%: 77.26% -> 77.87% | Improved |
| add non-linear projection head to contrastive learning (ablation) | AASN · Pelvic X-ray fracture detection (5-fold CV) | AUC: 97.31% -> 97.71% | Improved |
| add non-linear projection head to contrastive learning (ablation) | AASN · Pelvic X-ray fracture detection (5-fold CV) | Recall at FP=10%: 92.70% -> 92.71% | No clear effect |
| use AASN instead of CBN | CBN · Pelvic X-ray fracture detection (5-fold CV) | AUC: 97.00% -> 97.71% | Improved |
| use AASN instead of CBN | CBN · Pelvic X-ray fracture detection (5-fold CV) | AP: 94.92% -> 96.50% | Improved |
Closest papers by meaning
- Window Loss for Bone Fracture Detection and Localization in X-ray Images with Point-based Annotation (AAAI 2021 · similarity 0.86)
- Structured Landmark Detection via Topology-Adapting Deep Graph Learning (ECCV 2020 · similarity 0.85)
- Augmentation-invariant Learning Strategy via Data Augmentation for Improving Model Generalization (AAAI 2026 · similarity 0.83)
- Large-scale and Fine-grained Vision-language Pre-training for Enhanced CT Image Understanding (ICLR 2025 · similarity 0.83)
- Anatomical Knowledge Mining and Matching for Semi-supervised Medical Multi-structure Detection (AAAI 2025 · similarity 0.82)
- Dual-view X-ray Detection: Can AI Detect Prohibited Items from Dual-view X-ray Images like Humans? (CVPR 2025 · similarity 0.82)
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.
