Research map / Image generation / Image restoration
Unsupervised Domain-Specific Deblurring via Disentangled Representations
Boyu Lu, Jun-Cheng Chen, Rama Chellappa
Check 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 the blurring branch (ablation) | Unsupervised domain-specific deblurring model · CelebA test | PSNR: 18.83 -> 19.84 | Improved |
| add the blurring branch (ablation) | Unsupervised domain-specific deblurring model · CelebA test | d_VGG: 82.9 -> 65.5 | Improved |
| add content and blur disentanglement without KL regularization (ablation) | Unsupervised domain-specific deblurring model · CelebA test | PSNR: 19.84 -> 19.58 | Hurt |
| add the KL divergence loss to the blur encoder (ablation) | Unsupervised domain-specific deblurring model · CelebA test | PSNR: 19.58 -> 20.29 | Improved |
| add the KL divergence loss to the blur encoder (ablation) | Unsupervised domain-specific deblurring model · CelebA test | d_VGG: 69.8 -> 60.6 | Improved |
| add the perceptual loss (ablation) | Unsupervised domain-specific deblurring model · CelebA test | d_VGG: 60.6 -> 57.6 | Improved |
| use the perceptual loss | Unsupervised domain-specific deblurring model · Text deblurring | deblurring performance | Hurt |
| use the proposed unpaired method instead of the best reported baseline | BMVC_Text · BMVC_Text | PSNR: 30.6 -> 22.56 | Hurt |
| use the proposed unpaired method instead of the best reported baseline | BMVC_Text · BMVC_Text | CER: 7.2 -> 10.1 | Hurt |
| use the proposed method instead of the blurred-image baseline | Face verification · CFP F2F | accuracy: 0.920$\pm$0.014 -> 0.948$\pm$0.006 | Improved |
| use the proposed method instead of the blurred-image baseline | Face verification · CFP F2P | accuracy: 0.848$\pm$0.013 -> 0.872$\pm$0.015 | Improved |
Closest papers by meaning
- Unsupervised Class-Specific Deblurring (ECCV 2018 · similarity 0.90)
- Deep Semantic Face Deblurring (CVPR 2018 · similarity 0.89)
- Exploiting Semantics for Face Image Deblurring (IJCV 2020 · similarity 0.89)
- Blur2Blur: Blur Conversion for Unsupervised Image Deblurring on Unknown Domains (CVPR 2024 · similarity 0.89)
- Unsupervised Blind Image Deblurring Based on Self-Enhancement (CVPR 2024 · similarity 0.89)
- Cross-Scale Self-Supervised Blind Image Deblurring via Implicit Neural Representation (NeurIPS 2024 · 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.
