Research map / Point clouds / Shape and surface completion
Revealing the Invisible: Latent Structure Modeling for Semantically Consistent Cloud Removal
Jingwei Xin, Kai Guo, Jie Li, Nannan Wang
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 |
|---|---|---|---|
| add SR module alone (ablation) | VISER-CR baseline · CUHK-CR1 | PSNR: 25.62 -> 26.95 | Improved |
| add adaptive mask alone (ablation) | VISER-CR baseline · CUHK-CR1 | PSNR: 25.62 -> 28.75 | Improved |
| add patch saliency encoding on top of mask (ablation) | VISER-CR · CUHK-CR1 | PSNR: 28.75 -> 30.25 | Improved |
| add SR on top of PSE and mask (full model) (ablation) | VISER-CR · CUHK-CR1 | PSNR: 30.25 -> 31.11 | Improved |
| use masking ratio 0.25 instead of no masking (ablation) | VISER-CR · CUHK-CR1 | PSNR: 26.95 -> 31.11 | Improved |
| raise masking ratio from 0.25 to 0.50 (ablation) | VISER-CR · CUHK-CR1 | PSNR: 31.11 -> 26.82 | Hurt |
| raise masking ratio from 0.25 to 0.75 (ablation) | VISER-CR · CUHK-CR1 | PSNR: 31.11 -> 24.56 | Hurt |
| raise masking ratio from 0.25 to 0.35 (ablation) | VISER-CR · CUHK-CR1 | PSNR: 31.11 -> 29.03 | Hurt |
| increase temporal sequence length from 1 to 3 frames (ablation) | VISER-CR · Sen2 MTC New | PSNR: 16.17 -> 20.54 | Improved |
| increase temporal sequence length from 1 to 3 frames (ablation) | VISER-CR · Sen2 MTC New | LPIPS: 0.424 -> 0.231 | Improved |
| use VISER-CR instead of EMRDM | VISER-CR · CUHK-CR1 | PSNR: 27.281 -> 31.105 | Improved |
| use VISER-CR instead of EMRDM | VISER-CR · Sen2 MTC New | PSNR: 20.067 -> 20.541 | Improved |
Closest papers by meaning
- Interpretation-Oriented Cloud Removal via Observation-Anchored Residual Flow with Geo-Contextual Alignment (ECCV 2026 · similarity 0.90)
- Generative Adversarial Training for Weakly Supervised Cloud Matting (ICCV 2019 · similarity 0.86)
- Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusion (CVPR 2021 · similarity 0.86)
- Effective Cloud Removal for Remote Sensing Images by an Improved Mean-Reverting Denoising Model with Elucidated Design Space (CVPR 2025 · similarity 0.85)
- PIU-CR: Physics-Informed Deep Unfolding Network with SAR-Optical Image Fusion for Cloud Removal (NeurIPS 2026 · similarity 0.85)
- AllClear: A Comprehensive Dataset and Benchmark for Cloud Removal in Satellite Imagery (NeurIPS 2024 · similarity 0.83)
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.
