Research map / Label-efficient and robust learning / Noisy labels
Distributional Consistency Loss: Beyond Pointwise Data Terms in Inverse Problems
George Webber, Andrew J. Reader
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 MSE with distributional consistency loss in DIP denoising | Deep Image Prior · DIP Gaussian denoising (sigma 75/255) | PSNR over iterations | Improved |
| replace MSE with DC loss in DIP, compared against optimally early-stopped MSE | Deep Image Prior · DIP denoising across noise levels | peak PSNR | Improved |
| replace NLL with DC loss in Adam-based PET reconstruction | PET reconstruction (BrainWeb phantom) · 2D PET, Poisson noise, 10,000 iterations | image error over iterations | Improved |
| compare DC-Adam with MLEM at minimum error | PET reconstruction · 2D PET, Poisson noise | minimum error | Hurt |
| add TV regularization to DC loss versus NLL+TV (ablation) | PET reconstruction, 4x lower counts · 2D PET with edge-preserving TV | NRMSE | Improved |
| use DC loss so optimal TV strength drops (ablation) | PET with TV · 2D PET with edge-preserving TV | optimal regularization strength beta | Improved |
| apply DC loss to clinical 3D PET data | Real PET brain scan (Siemens Biograph mMR) · real 3D PET, 10,000 iterations, 21 subsets | late-iteration reconstruction quality | Improved |
| naive PIT for discrete noise instead of randomized PIT (ablation) | DC loss for Poisson noise · PET reconstruction | practicality | No clear effect |
Closest papers by meaning
- Provable Posterior Sampling with Denoising Oracles via Tilted Transport (NeurIPS 2024 · similarity 0.87)
- Solving Inverse Problems with Latent Diffusion Models via Hard Data Consistency (ICLR 2024 · similarity 0.87)
- Exact Posterior Score Estimation for Solving Linear Inverse Problems (NeurIPS 2026 · similarity 0.87)
- Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstruction (NeurIPS 2024 · similarity 0.87)
- Coupled Data and Measurement Space Dynamics for Enhanced Diffusion Posterior Sampling (NeurIPS 2025 · similarity 0.87)
- Improving Diffusion Posterior Samplers with Lagged Temporal Corrections for Image Restoration (NeurIPS 2026 · similarity 0.87)
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.
