Research map / Vision-language and multimodal / Human and embodied modeling
SynerMedGen: Synergizing Medical Multimodal Understanding with Generation via Task Alignment
Weiren zhao, DONG Yi, Cheng Chen
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 generation-aligned understanding training (stage I) without any generation training (ablation) | Bagel · BraTS T1→T2 | SSIM: 38.19 -> 79.91 | Improved |
| add generation-aligned understanding training (stage I) without any generation training (ablation) | Bagel · AutoPET PET→CT | SSIM: 12.19 -> 75.13 | Improved |
| add stage I generation-aligned understanding before stage II generation training (ablation) | SynerMedGen-UCG · Pelvis MRI→CT | SSIM: 73.99 -> 81.98 | Improved |
| add stage I generation-aligned understanding before stage II generation training (ablation) | SynerMedGen-UCG · BraTS FLAIR→T1 | SSIM: 81.34 -> 90.62 | Improved |
| replace generation-aligned understanding tasks with traditional medical understanding tasks in stage I (ablation) | SynerMedGen · 22 synthesis tasks (after stage I) | SSIM | Hurt |
| replace generation-aligned understanding tasks with traditional medical understanding tasks before stage II (ablation) | SynerMedGen · 22 synthesis tasks (after stage II) | SSIM | Hurt |
| add Conditional Target Selection task (ablation) | Bagel · 22 synthesis tasks | SSIM | Improved |
| add Modality Identification task on top of CTS (ablation) | SynerMedGen · 22 synthesis tasks | SSIM | Improved |
| add Transformation Instruction Alignment on top of CTS+MI (ablation) | SynerMedGen · 22 synthesis tasks | SSIM | Improved |
| add generation-aligned understanding training for unseen data | SynerMedGen-GAU vs Bagel/HealthGPT/UniMedVL · MyoPS (unseen cardiac MRI) | SSIM | Improved |
| add stage I understanding before generation training on unseen data (ablation) | SynerMedGen vs SynerMedGen-UCG · SynthRAD2025 (unseen) | SSIM | Improved |
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.
- UniMedVL — UniMedVL: Unifying Medical Multimodal Understanding and Generation through Observation-Knowledge-Analysis (ICML 2026) · better on 6 of 6 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
- SynerGen-VL: Towards Synergistic Image Understanding and Generation with Vision Experts and Token Folding (CVPR 2025 · similarity 0.85)
- RealUnify: Do Unified Models Truly Benefit from Unification? A Comprehensive Benchmark (CVPR 2026 · similarity 0.85)
- Generation Enhances Understanding in Unified Multimodal Models via Multi-Representation Generation (ICML 2026 · similarity 0.85)
- DuoGen: Towards Autonomous Interleaved Multimodal Generation (CVPR 2026 · similarity 0.85)
- Unison: Benchmarking Unified Multimodal Models via Synergistic Understanding and Generation (ICML 2026 · similarity 0.85)
- MedMax: Mixed-Modal Instruction Tuning for Training Biomedical Assistants (NeurIPS 2025 · similarity 0.85)
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.
