Research map / Image generation / Autoregressive and latent generation
DUAL-GLOW: Conditional Flow-Based Generative Model for Modality Transfer
Haoliang Sun, Ronak Mehta, Hao H. Zhou, Zhichun Huang, Sterling C. Johnson, Vivek Prabhakaran, Vikas Singh
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 |
|---|---|---|---|
| use DUAL-GLOW instead of C-VAE | MRI-to-PET generator · ADNI MRI-to-PET (10-fold CV, 806 pairs) | PSNR: 28.69 -> 29.56 | Improved |
| use DUAL-GLOW instead of C-VAE | MRI-to-PET generator · ADNI MRI-to-PET (10-fold CV, 806 pairs) | SSIM: 0.817 -> 0.898 | Improved |
| use DUAL-GLOW instead of C-VAE | MRI-to-PET generator · ADNI MRI-to-PET (10-fold CV, 806 pairs) | CorCoef: 0.980 -> 0.975 | Hurt |
| use DUAL-GLOW instead of pix2pix | MRI-to-PET generator · ADNI MRI-to-PET (10-fold CV, 806 pairs) | SSIM: 0.783 -> 0.898 | Improved |
| use synthetic PET instead of ground-truth PET for disease classification | SVM classifier · ADNI AD/CN SVM classification on 116 AAL ROI features | accuracy: 94% -> 91% | Hurt |
| use synthetic PET instead of ground-truth PET for disease classification | SVM classifier · ADNI AD/CN SVM classification on 116 AAL ROI features | false positive rate: 0% -> 3% | Hurt |
| use small lambda (0.001) on the MRI marginal likelihood regularizer (ablation) | DUAL-GLOW · ADNI MRI-to-PET | qualitative image quality | Improved |
| replace iterative Glow on concatenated MRI/PET with DUAL-GLOW | iGlow baseline · ADNI MRI-to-PET | qualitative image quality | Improved |
| condition on age with GRL-based exclusion discriminators | DUAL-GLOW · ADNI age conditioning | generated PET hypometabolism | Improved |
Closest papers by meaning
- CoCoLIT: ControlNet-Conditioned Latent Image Translation for MRI to Amyloid PET Synthesis (AAAI 2026 · similarity 0.87)
- Conditional Recurrent Flow: Conditional Generation of Longitudinal Samples With Applications to Neuroimaging (ICCV 2019 · similarity 0.86)
- A Simple Approach to Unifying Diffusion-based Conditional Generation (ICLR 2025 · similarity 0.85)
- DEFT: Efficient Fine-tuning of Diffusion Models by Learning the Generalised $h$-transform (NeurIPS 2024 · similarity 0.84)
- Toward a Vision-Language Foundation Model for Medical Data: Multimodal Dataset and Benchmarks for Vietnamese PET/CT Report Generation (NeurIPS 2025 · similarity 0.84)
- Discovering Latent Graphs with GFlowNets for Diverse Conditional Image Generation (NeurIPS 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.
