Research map / Video / Video segmentation and tracking
Multi-Scale Memory-Based Video Deblurring
Check 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 |
|---|---|---|---|
| Add the memory branch to the deblurring branch (ablation) | Video deblurring model · Downsampled GOPRO | PSNR: 30.88 -> 31.22 | Helped |
| Add bidirectional recurrence to the memory model (ablation) | Video deblurring model · Downsampled GOPRO | PSNR: 31.22 -> 31.63 | Helped |
| Add multi-scale processing to the memory model (ablation) | Video deblurring model · Downsampled GOPRO | PSNR: 31.22 -> 31.44 | Helped |
| Remove the deblurring branch and decode memory directly into a sharp frame (ablation) | Video deblurring model · Downsampled GOPRO | PSNR: 31.22 -> 30.58 | Hurt |
| Store the previous adjacent blurry-sharp pair as temporary memory (ablation) | Video deblurring model · Downsampled GOPRO | PSNR: 31.22 -> 31.03 | Hurt |
| Store h and r in separate memories under the same key instead of using one concatenated value (ablation) | Memory branch · Downsampled GOPRO | PSNR: 31.00 -> 31.22 | Helped |
| Memorize every frame rather than every 5 frames (ablation) | Video deblurring model, scale 1 · Downsampled GOPRO | PSNR: 31.22 -> 31.42 | Helped |
| Increase the memorization period from every frame to every 5 frames (ablation) | Video deblurring model, scale 1 · Downsampled GOPRO | Runtime: 0.582 -> 0.095 | Helped |
| Limit the memory bank to five recent frames (ablation) | Video deblurring model · GOPRO | Runtime: reduced the runtime to one-fifth | Helped |
| Limit the memory bank to five recent frames (ablation) | Video deblurring model · GOPRO | PSNR: 0.02dB | Hurt |
| Use the full bidirectional, multi-scale model instead of Ours (Slim) | Video deblurring model · Downsampled GOPRO | PSNR: +0.57dB | Helped |
| Use Ours rather than CDVD-TSP | Video deblurring · Original GOPRO | PSNR: 31.67 -> 31.76 | Helped |
Closest papers by meaning
- Cascaded Deep Video Deblurring Using Temporal Sharpness Prior (CVPR 2020 · similarity 0.89)
- Deep Video Deblurring for Hand-Held Cameras (CVPR 2017 · similarity 0.89)
- Deep Recurrent Neural Network with Multi-Scale Bi-directional Propagation for Video Deblurring (AAAI 2022 · similarity 0.88)
- Deep Multi-Scale Convolutional Neural Network for Dynamic Scene Deblurring (CVPR 2017 · similarity 0.87)
- Blur-aware Spatio-temporal Sparse Transformer for Video Deblurring (CVPR 2024 · similarity 0.85)
- Efficient Multi-scale Network with Learnable Discrete Wavelet Transform for Blind Motion Deblurring (CVPR 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.