Research map / Image generation / Autoregressive and latent generation
CoCoLIT: ControlNet-Conditioned Latent Image Translation for MRI to Amyloid PET Synthesis
Alec Sargood, Lemuel Puglisi, James H Cole, Neil P. Oxtoby, Daniele Ravì, Daniel C. Alexander
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 time-weighted image-space loss (WISL) to ControlNet training (ablation) | CoCoLIT · ADNI internal test | SSIM: 0.841 -> 0.865 | Improved |
| add constant-weight image-space loss (ISL) (ablation) | CoCoLIT · ADNI internal test | SSIM: 0.841 -> 0.870 | Improved |
| WISL vs constant-weight ISL when combined with LAS (ablation) | CoCoLIT · ADNI internal test | Balanced accuracy (%): 56.7% -> 62.3% | Improved |
| add Latent Average Stabilization at inference (ablation) | CoCoLIT · ADNI internal test | SSIM: 0.865 -> 0.896 | Improved |
| increase LAS samples m from 1 to 64 (ablation) | CoCoLIT · ADNI internal test | PSNR: 22.570 -> 24.135 | Improved |
| increase LAS samples m from 1 to 64 (ablation) | CoCoLIT · ADNI internal test | Balanced accuracy (%): 57.4% -> 62.3% | Mixed |
| decode averaged latent (LAS) instead of averaging 64 decoded images (ablation) | CoCoLIT · ADNI internal test | Balanced accuracy (%): 60.1% -> 62.3% | No clear effect |
| CoCoLIT vs best baseline on amyloid-positivity classification | MRI-to-PET synthesis · A4 external test | Balanced accuracy (%): 56.1% -> 79.8% | Improved |
| CoCoLIT vs best baseline on amyloid-positivity classification | MRI-to-PET synthesis · ADNI internal test | Balanced accuracy (%): 51.8% -> 62.3% | Improved |
| add WISL without LAS vs Base on balanced accuracy (ablation) | CoCoLIT · ADNI internal test | Balanced accuracy (%): 43.9% -> 57.4% | Improved |
| add LAS on top of WISL-trained model on SSIM (ablation) | CoCoLIT · ADNI internal test | HABC: 0.334 -> 0.522 | Mixed |
Closest papers by meaning
- DUAL-GLOW: Conditional Flow-Based Generative Model for Modality Transfer (ICCV 2019 · similarity 0.87)
- An Optimal Transportation Based Univariate Neuroimaging Index (ICCV 2017 · similarity 0.85)
- EMAD: Evidence-Centric Grounded Multimodal Diagnosis for Alzheimer's Disease (CVPR 2026 · similarity 0.84)
- Progression as Latent Drift: Generative Forecasting of Slow-Evolving Pathologies (ECCV 2026 · similarity 0.83)
- Simpler Diffusion: 1.5 FID on ImageNet512 with Pixel-space Diffusion (CVPR 2025 · similarity 0.83)
- Autoregressive Image Diffusion: Generation of Image Sequence and Application in MRI (NeurIPS 2024 · 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.
