Research map / Image generation / Image restoration
Burst Denoising With Kernel Prediction Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 per-pixel kernel prediction to jointly align and denoise burst frames | Burst denoising KPN · Synthetic linear test, Gain ∝ 1 | PSNR: 34.60 -> 36.47 | Helped |
| Add the per-pixel noise estimate as a network input (ablation) | Burst denoising KPN · Synthetic linear test, Gain ∝ 8 | PSNR: 22.37 -> 27.97 | Helped |
| Add the per-pixel noise estimate as a network input (ablation) | Burst denoising KPN · Synthetic AWGN test, PSNR = 10 | PSNR: 10.14 -> 23.89 | Helped |
| Add annealing with an intermediate per-frame loss (ablation) | Burst denoising KPN · Synthetic linear test, Gain ∝ 4 | PSNR: 30.46 -> 31.19 | Helped |
| Use all burst frames instead of a single frame (ablation) | Burst denoising KPN · Synthetic linear test, Gain ∝ 4 | PSNR: 29.22 -> 31.19 | Helped |
| Predict kernels rather than synthesize output pixels directly (ablation) | Burst denoising KPN · Synthetic AWGN test, PSNR = 10 | PSNR: 25.43 -> 23.89 | Hurt |
| Predict kernels rather than synthesize output pixels directly (ablation) | Burst denoising KPN · Synthetic linear test, Gain ∝ 1 | PSNR: 35.93 -> 36.47 | Helped |
| Predict kernels rather than synthesize output pixels directly (ablation) | Burst denoising KPN · Synthetic linear test, Gain ∝ 8 | SSIM: 0.748 -> 0.741 | Hurt |
Closest papers by meaning
- Basis Prediction Networks for Effective Burst Denoising With Large Kernels (CVPR 2020 · similarity 0.90)
- Iterative Residual CNNs for Burst Photography Applications (CVPR 2019 · similarity 0.88)
- NERDS: A General Framework to Train Camera Denoisers from Raw-RGB Noisy Image Pairs (ICLR 2023 · similarity 0.88)
- Universal Denoising Networks : A Novel CNN Architecture for Image Denoising (CVPR 2018 · similarity 0.87)
- Classic Video Denoising in a Machine Learning World: Robust, Fast, and Controllable (CVPR 2025 · similarity 0.86)
- Zero-Shot Noise2Mean: Gap Minimization for Efficient Denoising from a Single Noisy Image (AAAI 2025 · similarity 0.86)
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.