Research map / Video / Video generation
BokehCrafter: Taming Video Diffusion Models for Controllable Bokeh Rendering
Qiwen Wang, Liao Shen, Jiaqi Li, Tianqi Liu, Huiqiang Sun, Zihao Huang, Yachuan Huang, Xianrui Luo, Zhiguo Cao
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 |
|---|---|---|---|
| select reference frame from a different scene (Reference Content Decoupling) (ablation) | BokehCrafter · VBS test | PSNR: 26.70 -> 27.86 | Improved |
| add Bokeh Image Extraction (Q-Former) on reference image features (ablation) | BokehCrafter · VBS test | PSNR: 23.31 -> 27.86 | Improved |
| add reference image stream via Text-Image Fusion (ablation) | BokehCrafter · VBS test | PSNR: 26.93 -> 27.86 | Improved |
| add BIE (ablation) | BokehCrafter · VBS test | LPIPS: 0.3527 -> 0.1219 | Improved |
| video diffusion backbone vs BokehMe/VBR | BokehCrafter · VBS test, large blur | PSNR: 26.52 -> 26.70 (vs BokehMe) | Improved |
| video diffusion backbone vs VBR | BokehCrafter · VBS test, medium blur | PSNR: 28.01 -> 28.35 (vs VBR) | Improved |
| video diffusion with temporal attention | BokehCrafter · VBS test, small blur | temporal consistency error: 0.0385 -> 0.0352 (vs BokehMe) | Improved |
| use BokehCrafter instead of VBR | BokehCrafter · User study | human preference: 62.35% preferred ours | Improved |
Closest papers by meaning
- Dr. Bokeh: DiffeRentiable Occlusion-aware Bokeh Rendering (CVPR 2024 · similarity 0.90)
- Diffusion Renderer: Neural Inverse and Forward Rendering with Video Diffusion Models (CVPR 2025 · similarity 0.87)
- AnyBokeh: Physics-Guided Any-to-Any Bokeh Editing with Optical Fingerprint Transfer (NeurIPS 2026 · similarity 0.86)
- BLURD: Benchmarking and Learning using a Unified Rendering and Diffusion Model (NeurIPS 2024 · similarity 0.86)
- ViBiDSampler: Enhancing Video Interpolation Using Bidirectional Diffusion Sampler (ICLR 2025 · similarity 0.86)
- VideoElevator: Elevating Video Generation Quality with Versatile Text-to-Image Diffusion Models (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.
