Research map / Probabilistic and scientific modeling / Causal discovery
Towards Ultrasound-based Reliable Disease Diagnosis Using Causal Inference
Bolei Chen, Jiaxu Kang, Haonan Yang, Ping Zhong, Yixiong Liang, Rui Fan, Jianxin Wang
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 |
|---|---|---|---|
| replace vanilla spatial attention with CA module (ablation) | CEUS Transformer branch · BUS-CEUS | Accuracy (%): 79.25 -> 83.67 | Improved |
| add B-Down and B-Up feature interaction bridges without causal modules (ablation) | dual-branch model · BUS-CEUS | Accuracy (%): 83.67 -> 85.03 | Improved |
| add FACL to bridges (ablation) | dual-branch model · BUS-CEUS | Accuracy (%): 85.03 -> 88.37 | Improved |
| add BACL to bridges (instead of FACL) (ablation) | dual-branch model · BUS-CEUS | Accuracy (%): 85.03 -> 89.07 | Improved |
| add both BACL and FACL (ablation) | dual-branch model · BUS-CEUS | Accuracy (%): 89.07 -> 89.80 | Improved |
| increase CEUS frames k or image size beyond k=7 and 384x384 (ablation) | full model · BUS-CEUS | diagnostic performance | Hurt |
| use causal dual-branch model instead of UAC-T | multimodal diagnosis · BUS-CEUS | Accuracy (%): 86.33 -> 89.80 | Improved |
| use method instead of Li et al. on unimodal BUSI | BUSI · BUSI | Sensitivity (%): 90.0 -> 89.8 sens; 90.2 -> 91.6 acc | Mixed |
| compare compute with UAC-T and Trans-CEUS | Ours · BUS-CEUS | Params / FLOPs: 63.41M params, 48.28G FLOPs vs UAC-T 40.55M, 67.34G | Mixed |
Closest papers by meaning
- Learning Causal Alignment for Reliable Disease Diagnosis (ICLR 2025 · similarity 0.87)
- U2-BENCH: Benchmarking Large Vision-Language Models on Ultrasound Understanding (ICLR 2026 · similarity 0.85)
- The Boundaries of Fair AI in Medical Image Prognosis: A Causal Perspective (NeurIPS 2025 · similarity 0.84)
- RAD: Towards Trustworthy Retrieval-Augmented Multi-modal Clinical Diagnosis (NeurIPS 2025 · similarity 0.83)
- Multi-modal Medical Diagnosis via Large-small Model Collaboration (CVPR 2025 · similarity 0.83)
- PersonalizedUS: Interpretable Breast Cancer Risk Assessment with Local Coverage Uncertainty Quantification (AAAI 2025 · similarity 0.83)
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.
