Research map / Image generation / Image restoration
Deep Image Prior
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 |
|---|---|---|---|
| average restored images from the last optimization iterations (ablation) | Deep Image Prior · Standard denoising dataset, σ = 25 | PSNR: 29.22 -> 30.43 | Helped |
| average the results of two optimization runs (ablation) | Deep Image Prior · Standard denoising dataset, σ = 25 | PSNR: 30.43 -> 31.00 | Helped |
| use Deep Image Prior instead of bicubic upsampling | 4x super-resolution · Set5 | PSNR: 28.43 -> 29.90 | Helped |
| use Deep Image Prior instead of bicubic upsampling | 4x super-resolution · Set14 | PSNR: 26.05 -> 27.00 | Helped |
| use Deep Image Prior instead of trained SRResNet | 4x super-resolution · Set5 | PSNR: 32.10 -> 29.90 | Hurt |
| use Deep Image Prior instead of trained SRResNet | 4x super-resolution · Set14 | PSNR: 28.53 -> 27.00 | Hurt |
| use Deep Image Prior instead of Papyan et al.'s convolutional sparse coding | Inpainting · Barbara | PSNR: 28.14 -> 32.22 | Helped |
| use deeper encoder-decoder architectures (ablation) | Deep Image Prior inpainting · Inpainting | visual restoration quality | Helped |
| add U-Net skip connections to the ResNet architecture (ablation) | Deep Image Prior inpainting · Inpainting | visual restoration quality | Hurt |
| increase optimization from 2400 to 50K iterations (ablation) | Deep Image Prior JPEG restoration · JPEG-compressed image restoration | visual restoration quality | Hurt |
Closest papers by meaning
- A Restoration Network as an Implicit Prior (ICLR 2024 · similarity 0.90)
- Denoising and Regularization via Exploiting the Structural Bias of Convolutional Generators (ICLR 2020 · similarity 0.89)
- Stochastic Deep Restoration Priors for Imaging Inverse Problems (ICML 2025 · similarity 0.88)
- Omni-Kernel Network for Image Restoration (AAAI 2024 · similarity 0.87)
- A Principled Self-Referenced Early Stopping Approach for Deep Image Prior (NeurIPS 2026 · similarity 0.87)
- UP-Restorer: When Unrolling Meets Prompts for Unified Image Restoration (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.