Research map / Image generation / Autoregressive and latent generation
Deep Image Matting
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 compositional loss to alpha prediction loss (ablation) | Encoder-Decoder network · Composition-1k | SAD: 59.6 -> 54.6 | Helped |
| add compositional loss to alpha prediction loss (ablation) | Encoder-Decoder network · Composition-1k | MSE: 0.019 -> 0.017 | Helped |
| add compositional loss to alpha prediction loss (ablation) | Encoder-Decoder network · Composition-1k | Gradient: 40.5 -> 36.7 | Helped |
| add compositional loss to alpha prediction loss (ablation) | Encoder-Decoder network · Composition-1k | Connectivity: 59.3 -> 55.3 | Helped |
| add a trained refinement network after the encoder-decoder (ablation) | Encoder-Decoder network · Composition-1k | SAD: 54.6 -> 50.4 | Helped |
| add a trained refinement network after the encoder-decoder (ablation) | Encoder-Decoder network · Composition-1k | MSE: 0.017 -> 0.014 | Helped |
| add a trained refinement network after the encoder-decoder (ablation) | Encoder-Decoder network · Composition-1k | Gradient: 36.7 -> 31.0 | Hurt |
| add a trained refinement network after the encoder-decoder (ablation) | Encoder-Decoder network · Composition-1k | Connectivity: 55.3 -> 50.8 | Helped |
| post-process encoder-decoder output with Guided filter (ablation) | Encoder-Decoder network · Composition-1k | SAD: 54.6 -> 52.2 | Helped |
| post-process encoder-decoder output with Guided filter (ablation) | Encoder-Decoder network · Composition-1k | MSE: 0.017 -> 0.016 | Helped |
| post-process encoder-decoder output with Guided filter (ablation) | Encoder-Decoder network · Composition-1k | Gradient: 36.7 -> 30.0 | Helped |
| post-process encoder-decoder output with Guided filter (ablation) | Encoder-Decoder network · Composition-1k | Connectivity: 55.3 -> 52.6 | Helped |
Closest papers by meaning
- A Late Fusion CNN for Digital Matting (CVPR 2019 · similarity 0.93)
- High-Resolution Deep Image Matting (AAAI 2021 · similarity 0.91)
- Training Matting Models Without Alpha Labels (AAAI 2025 · similarity 0.91)
- Revisiting Context Aggregation for Image Matting (ICML 2024 · similarity 0.88)
- In-Context Matting (CVPR 2024 · similarity 0.88)
- DRIP: Unleashing Diffusion Priors for Joint Foreground and Alpha Prediction in Image Matting (NeurIPS 2024 · similarity 0.88)
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.