Research map / Learning theory and algorithms / Matrix and tensor methods
Momentum-Net: Fast and Convergent Iterative Neural Network for Inverse Problems
Il Yong Chun, Zhengyu Huang, Hongki Lim, Jeffrey A. Fessler
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 extrapolation (momentum) module to Momentum-Net (ablation) | Momentum-Net (sCNN) · Sparse-view CT (to reach 24 HU RMSE) | MBIR time reduction (%): 53.3% less MBIR time | Improved |
| add extrapolation (momentum) module to Momentum-Net (ablation) | Momentum-Net (sCNN) · Light-field photography from focal stack (to reach 32 dB PSNR) | MBIR time reduction (%): 36.5% less MBIR time | Improved |
| use Momentum-Net with noniterative MBIR module instead of BCD-Net with 3 inner iterations | BCD-Net · Sparse-view CT (to reach 24 HU RMSE) | MBIR time reduction (%): 62.5% less MBIR time | Improved |
| use Momentum-Net with noniterative MBIR module instead of BCD-Net with 3 inner iterations | BCD-Net · Light-field photography from focal stack (to reach 32 dB PSNR) | MBIR time reduction (%): 61.5% less MBIR time | Improved |
| use deep CNN (dCNN) refiners instead of shallow CNN (sCNN) refiners (ablation) | Momentum-Net · Sparse-view CT and light-field photography | total run time | No clear effect |
| use L=4 dCNN layers instead of L=6 with R=3^3 (ablation) | Momentum-Net (dCNN) · Sparse-view CT | RMSE | Improved |
| use Momentum-Net instead of MBIR with learned convolutional regularizer | MBIR with learned convolutional regularizer · Sparse-view CT (12.5% views) | reconstruction quality | Improved |
| use Momentum-Net instead of MBIR with 4D EP regularizer | 4D EP MBIR · Light-field photography from focal stack | PSNR / depth estimation accuracy | Improved |
Closest papers by meaning
- An analytic theory of convolutional neural network inverse problems solvers (ICML 2026 · similarity 0.86)
- NPN: Non-Linear Projections of the Null-Space for Imaging Inverse Problems (NeurIPS 2025 · similarity 0.84)
- A Unified Framework for Forward and Inverse Problems in Subsurface Imaging using Latent Space Translations (ICLR 2025 · similarity 0.84)
- Weakly Convex Regularisers for Inverse Problems: Convergence of Critical Points and Primal-Dual Optimisation (ICML 2024 · similarity 0.84)
- Integrating Intermediate Layer Optimization and Projected Gradient Descent for Solving Inverse Problems with Diffusion Models (ICML 2025 · similarity 0.84)
- A Restoration Network as an Implicit Prior (ICLR 2024 · similarity 0.84)
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.
