Research map / Image generation / Autoregressive and latent generation
dugMatting: Decomposed-Uncertainty-Guided Matting
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 |
|---|---|---|---|
| Apply dugMatting without user interaction | P3MNet · P3M-500-P | SAD: 12.73 -> 10.08 | Helped |
| Apply dugMatting without user interaction | P3MNet · P3M-500-NP | SAD: 16.49 -> 10.30 | Helped |
| Add user-map interaction to dugMatting | P3MNet · P3M-500-P | SAD: 10.08 -> 7.72 | Helped |
| Add user-map interaction to dugMatting | P3MNet · P3M-500-NP | SAD: 10.30 -> 8.79 | Helped |
| Apply dugMatting with one or two interaction selections | MatteFormer · Composition-1K | SAD (10^3): 23.8 -> 25.8; 23.8 -> 23.4 | Mixed |
| Add interaction selections (ablation) | MatteFormer (dugMatting) · Composition-1K | SAD (10^3): 34.1 -> 25.8 -> 23.4 | Helped |
| Replace deterministic output with an NIG distribution (ablation) | MODNet · P3M-500-P | SAD: 23.86 -> 20.04 | Helped |
| Add the refinement module to NIG output (ablation) | MODNet · P3M-500-P | SAD: 20.04 -> 18.15 | Helped |
| Replace deterministic output with an NIG distribution (ablation) | SHM · P3M-500-P | SAD: 26.84 -> 24.65 | Helped |
| Add the refinement module to NIG output (ablation) | SHM · P3M-500-P | SAD: 24.65 -> 21.43 | Helped |
| Use pixel-specific uncertainty-guided Gaussian augmentation instead of fixed-variance Gaussian noise (ablation) | Refinement module · P3M-500-P | SAD_f: 3.36 -> 3.36 | No clear effect |
| Use pixel-specific uncertainty-guided Gaussian augmentation instead of fixed-variance Gaussian noise (ablation) | Refinement module · P3M-500-P | SAD_b: 6.55 -> 6.23 | Helped |
Closest papers by meaning
- Unifying Automatic and Interactive Matting with Pretrained ViTs (CVPR 2024 · similarity 0.88)
- MP-Mat: A 3D-and-Instance-Aware Human Matting and Editing Framework with Multiplane Representation (ICLR 2025 · similarity 0.86)
- Training Matting Models Without Alpha Labels (AAAI 2025 · similarity 0.86)
- In-Context Matting (CVPR 2024 · similarity 0.86)
- MatAnyone: Stable Video Matting with Consistent Memory Propagation (CVPR 2025 · similarity 0.86)
- Privileged Prior Information Distillation for Image Matting (AAAI 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.