Research map / Detection and segmentation / CNN features
A Late Fusion CNN for Digital 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 |
|---|---|---|---|
| Use a two-decoder late-fusion network without trimap input | Late fusion CNN · Human image matting testing dataset, transition regions | SAD: 10.39 -> 10.08 | Helped |
| Use a two-decoder late-fusion network without trimap input | Late fusion CNN · Composition-1k testing dataset, transition regions | SAD: 33.64 -> 49.05 | Hurt |
| Add the background and fusion branches to the segmentation-only stage (ablation) | Late fusion CNN · Human image matting testing dataset, transition regions | SAD: 20.93 -> 10.08 | Helped |
| Add the background and fusion branches to the segmentation-only stage (ablation) | Late fusion CNN · Composition-1k testing dataset, transition regions | SAD: 103.21 -> 49.05 | Helped |
| Jointly train the segmentation and fusion networks rather than use the fusion-pretraining stage alone (ablation) | Late fusion CNN · Human image matting testing dataset, transition regions | SAD: 14.23 -> 10.08 | Helped |
| Jointly train the segmentation and fusion networks rather than use the fusion-pretraining stage alone (ablation) | Late fusion CNN · Composition-1k testing dataset, transition regions | SAD: 66.05 -> 49.05 | Helped |
| Apply guided filter refinement (ablation) | Late fusion CNN · Human image matting testing dataset, transition regions | SAD: 10.08 -> 9.75 | Helped |
| Apply guided filter refinement (ablation) | Late fusion CNN · Composition-1k testing dataset, transition regions | SAD: 49.05 -> 49.02 | No clear effect |
Closest papers by meaning
- Deep Image Matting (CVPR 2017 · similarity 0.93)
- Attention-Guided Hierarchical Structure Aggregation for Image Matting (CVPR 2020 · similarity 0.90)
- Training Matting Models Without Alpha Labels (AAAI 2025 · similarity 0.90)
- DRIP: Unleashing Diffusion Priors for Joint Foreground and Alpha Prediction in Image Matting (NeurIPS 2024 · similarity 0.87)
- Revisiting Context Aggregation for Image Matting (ICML 2024 · similarity 0.87)
- In-Context Matting (CVPR 2024 · similarity 0.87)
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.