Research map / Point clouds / Segmentation and detection
Generative Adversarial Training for Weakly Supervised Cloud Matting
Zhengxia Zou, Wenyuan Li, Tianyang Shi, Zhenwei Shi, Jieping Ye
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 adversarial training (ablation) | Ours LSGAN+UNet · GF-1 cloud detection | AP avg: 0.9335 -> 0.9523 | Improved |
| add saturation penalty to generator objective (ablation) | Ours LSGAN+UNet · GF-1 cloud detection | training convergence | Improved |
| add end-to-end training of G, D and F (ablation) | Ours LSGAN+UNet · GF-1 cloud detection | AP avg: 0.9430 -> 0.9523 | Improved |
| add background image as generator input (ablation) | Ours LSGAN+UNet · GF-1 cloud detection | AP avg: 0.9483 -> 0.9523 | Improved |
| use LSGAN instead of vanilla GAN | Ours +UNet · GF-1 cloud detection | AP avg: 0.9462 -> 0.9523 | Improved |
| use LSGAN instead of vanilla GAN | Ours +UNet · GF-1 cloud removal | MAE: 0.0720 -> 0.0753 | Hurt |
| replace UNet encoder with VGG | Ours LSGAN · GF-1 cloud detection | AP avg: 0.9523 -> 0.9538 | Improved |
| replace UNet encoder with Resnet50 | Ours LSGAN · GF-1 cloud detection | AP avg: 0.9523 -> 0.9526 | No clear effect |
| compare weakly supervised method with fully supervised UNet | Ours LSGAN+UNet · GF-1 cloud detection | AP avg: 0.9613 -> 0.9523 | Hurt |
| use proposed method instead of CycleGAN for cloud removal | Ours LSGAN+UNet · GF-1 cloud removal | MAE: 0.0904 -> 0.0753 | Improved |
| add cloud augmentation to training | SSD (VGG) · Occluded airplane detection | VOC2012-AP: 78.9% -> 81.7% | Improved |
| add cloud augmentation to training | RetinaNet (Resnet-50) · Occluded airplane detection | VOC2012-AP: 83.3% -> 87.3% | Improved |
Closest papers by meaning
- Revealing the Invisible: Latent Structure Modeling for Semantically Consistent Cloud Removal (AAAI 2026 · similarity 0.86)
- Unsupervised Point Cloud Completion and Segmentation by Generative Adversarial Autoencoding Network (NeurIPS 2022 · similarity 0.85)
- Weakly Supervised Semantic Segmentation for Large-Scale Point Cloud (AAAI 2021 · similarity 0.84)
- Weakly Supervised Segmentation on Outdoor 4D Point Clouds With Temporal Matching and Spatial Graph Propagation (CVPR 2022 · similarity 0.84)
- Preface to the Special Issue on Pattern Recognition (DAGM GCPR 2021) (IJCV 2023 · similarity 0.84)
- Weakly Supervised Semantic Point Cloud Segmentation: Towards 10x Fewer Labels (CVPR 2020 · similarity 0.84)
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.
