Research map / Detection and segmentation / CNN features
Instance Shadow Detection with A Single-Stage Detector
Tianyu Wang, Xiaowei Hu, Pheng‐Ann Heng, Chi‐Wing Fu
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 deformable MaskIoU head (ablation) | SSIS · SOBA-testing | SOAP segm: 30.2 -> 32.0 | Improved |
| add boundary loss (ablation) | SSIS · SOBA-testing | SOAP segm: 30.2 -> 31.0 | Improved |
| add shadow-aware copy-and-paste augmentation (ablation) | SSIS · SOBA-testing | SOAP segm: 30.2 -> 31.1 | Improved |
| add all three new components (ablation) | SSIS · SOBA-testing | SOAP segm: 30.2 -> 35.3 | Improved |
| add offset vectors (ablation) | basic network · SOBA-testing | SOAP segm: 28.1 -> 29.1 | Improved |
| add class vectors (ablation) | basic + offset · SOBA-testing | SOAP segm: 29.1 -> 30.2 | Improved |
| add MaskIoU head (ablation) | SSISv2 · SOBA-testing | SOAP segm: 32.1 -> 35.3 | Improved |
| use deformable conv in MaskIoU head (ablation) | SSISv2 · SOBA-testing | SOAP bbox: 25.5 -> 29.0 | Improved |
| add boundary loss (thick and thin) (ablation) | SSISv2 · SOBA-testing | SOAP segm: 33.5 -> 35.3 | Improved |
| add shadow-aware copy-and-paste with shadows (ablation) | SSISv2 · SOBA-testing | SOAP segm: 33.3 -> 35.3 | Improved |
| copy-paste object only without shadows (ablation) | SSISv2 · SOBA-testing | SOAP segm: 33.3 -> 32.8 | Hurt |
| learn both directions with SSIS inference instead of object to shadow only (ablation) | SSIS · SOBA-testing | SOAP segm: 23.8 -> 30.2 | Improved |
Closest papers by meaning
- Instance Shadow Detection (CVPR 2020 · similarity 0.94)
- Single-Stage Instance Shadow Detection With Bidirectional Relation Learning (CVPR 2021 · similarity 0.93)
- Cast and Attached Shadow Detection via Iterative Light and Geometry Reasoning (ECCV 2026 · similarity 0.89)
- MetaShadow: Object-Centered Shadow Detection, Removal, and Synthesis (CVPR 2025 · similarity 0.86)
- Shadow-Enlightened Image Outpainting (CVPR 2024 · similarity 0.84)
- DeS3: Adaptive Attention-Driven Self and Soft Shadow Removal Using ViT Similarity (AAAI 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.
