Research map / Image generation / Fast diffusion sampling
DOLCE: A Model-Based Probabilistic Diffusion Framework for Limited-Angle CT Reconstruction
Jiaming Liu, Rushil Anirudh, Jayaraman J. Thiagarajan, Stewart He, K Aditya Mohan, Ulugbek S. Kamilov, Hyojin Kim
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 |
|---|---|---|---|
| condition the diffusion model on limited-angle reconstruction and use DOLCE instead of unconditionally trained diffusion (DPS) | DOLCE · Body CT, 60 deg | PSNR (dB): 28.97 -> 35.11 | Improved |
| add proximal data-consistency step during sampling (ablation) | DOLCE (RLS condition) · Body CT, 120 deg | PSNR (dB): 41.25 -> 42.52 | Improved |
| add proximal data-consistency step during sampling (ablation) | DOLCE (RLS condition) · Luggage CT, 90 deg | PSNR (dB): 38.64 -> 39.28 | Improved |
| use RLS instead of FBP as the conditional input (ablation) | DOLCE · Body CT, 60 deg | PSNR (dB): 34.05 -> 35.15 | Improved |
| average multiple samples (DOLCE-SA) (ablation) | DOLCE · Body CT, 60 deg | PSNR (dB): 35.11 -> 35.58 | Improved |
| train one DOLCE model jointly on luggage and medical data instead of separate models (ablation) | DOLCE · Luggage and body CT, 60 to 120 deg | PSNR/SSIM: 33.59 -> 33.98 (luggage 60 deg) | No clear effect |
| test on a different distribution than trained (model trained on medical, tested on luggage) (ablation) | DOLCE · Luggage CT, 60 deg | PSNR (dB): 33.59 -> 26.78 | Hurt |
| replace uniform schedule with cosine noise schedule during training (ablation) | DOLCE · CT reconstruction | image reconstruction quality | No clear effect |
| test with unseen angular range (45 deg) not seen in training | DOLCE · Body CT, 45 deg | PSNR (dB): 24.50 -> 24.97 (DPS vs DOLCE) | Improved |
| increase number of sampling steps K (ablation) | DOLCE · Body CT | PSNR | 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.
- CTNet — CT-Net: Complementary Transfering Network for Garment Transfer With Arbitrary Geometric Changes (CVPR 2021) · better on 1 of 1 shared result
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
- DEFT: Efficient Fine-tuning of Diffusion Models by Learning the Generalised $h$-transform (NeurIPS 2024 · similarity 0.87)
- Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstruction (NeurIPS 2024 · similarity 0.86)
- DiffusionBlend: Learning 3D Image Prior through Position-aware Diffusion Score Blending for 3D Computed Tomography Reconstruction (NeurIPS 2024 · similarity 0.86)
- A Variational Perspective on Solving Inverse Problems with Diffusion Models (ICLR 2024 · similarity 0.86)
- Diffusion Priors for Variational Likelihood Estimation and Image Denoising (NeurIPS 2024 · similarity 0.86)
- Principled Probabilistic Imaging using Diffusion Models as Plug-and-Play Priors (NeurIPS 2024 · 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.
