Research map / Probabilistic and scientific modeling / Variational and generative sampling
UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders
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 CVAE to model annotation variants rather than deterministic point predictions (ablation) | RGB-D saliency detection · NJU2K | Fβ: .851 -> .886 | Helped |
| Use CVAE to model annotation variants rather than deterministic point predictions (ablation) | RGB-D saliency detection · SSB | Fβ: .834 -> .884 | Helped |
| Use CVAE to model annotation variants rather than deterministic point predictions (ablation) | RGB-D saliency detection · DES | Fβ: .867 -> .919 | Helped |
| Use CVAE to model annotation variants rather than deterministic point predictions (ablation) | RGB-D saliency detection · NLPR | Fβ: .845 -> .891 | Helped |
| Use CVAE to model annotation variants rather than deterministic point predictions (ablation) | RGB-D saliency detection · LFSD | Fβ: .802 -> .855 | Helped |
| Use CVAE to model annotation variants rather than deterministic point predictions (ablation) | RGB-D saliency detection · SIP | Fβ: .821 -> .867 | Helped |
| Use latent dimension K=8 instead of K=32 (ablation) | UC-Net · NJU2K | Fβ: .858 -> .886 | Helped |
| Add DepthCorrectionNet to refine noisy depth (ablation) | UC-Net · DES | Fβ: .868 -> .919 | Helped |
| Add saliency consensus over multiple sampled predictions (ablation) | UC-Net · NJU2K | Fβ: .884 -> .886 | Helped |
| Replace the CVAE with a multi-head model (ablation) | UC-Net · DES | Fβ: .897 -> .919 | Hurt |
| Replace the CVAE with Monte-Carlo dropout (ablation) | UC-Net · DES | Fβ: .868 -> .919 | Hurt |
| Replace raw depth with HHA depth encoding (ablation) | UC-Net · NJU2K | Fβ: .884 -> .886 | No clear effect |
Closest papers by meaning
- Promoting Saliency From Depth: Deep Unsupervised RGB-D Saliency Detection (ICLR 2022 · similarity 0.89)
- Select, Supplement and Focus for RGB-D Saliency Detection (CVPR 2020 · similarity 0.89)
- Energy-Based Generative Cooperative Saliency Prediction (AAAI 2022 · similarity 0.88)
- Learning Selective Self-Mutual Attention for RGB-D Saliency Detection (CVPR 2020 · similarity 0.88)
- Alignment-Free RGB-T Salient Object Detection: A Large-Scale Dataset and Progressive Correlation Network (AAAI 2025 · similarity 0.85)
- Advancing Saliency Ranking with Human Fixations: Dataset Models and Benchmarks (CVPR 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.