Research map / Detection and segmentation / CNN features
Single Image Water Hazard Detection using FCN with Reflection Attention Units
Xiaofeng Han, Chuong Nguyen, Shaodi You, Jianfeng Lu
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 5 RAUs instead of 5 plain conv layers (ablation) | FCN-8s-FL · ONR | F-measure (%): 55.76% -> 70.11% | Improved |
| add 5 RAUs instead of 5 plain conv layers (ablation) | FCN-8s-FL · OFR | F-measure (%): 78.56% -> 81.67% | Improved |
| add 5 RAUs instead of 5 plain conv layers (ablation) | FCN-8s-FL · BOTH (ONR+OFR) | F-measure (%): 63.44% -> 76.91% | Improved |
| increase RAUs from 1 to 5 (ablation) | FCN-8s-FL · ONR | F-measure (%): 59.63% -> 70.11% | Improved |
| increase RAUs from 3 to 5 (ablation) | FCN-8s-FL · OFR | F-measure (%): 79.09% -> 81.67% | Improved |
| increase RAUs from 3 to 5 (ablation) | FCN-8s-FL · BOTH (ONR+OFR) | F-measure (%): 67.63% -> 76.91% | Improved |
| add focal loss (ablation) | FCN-8s · OFR | F-measure (%): 64.33% -> 74.05% | Improved |
| add focal loss (ablation) | FCN-8s · BOTH (ONR+OFR) | F-measure (%): 65.21% -> 70.62% | Improved |
| add focal loss (ablation) | FCN-8s · ONR | F-measure (%): 56.99% -> 57.85% | No clear effect |
| add RAUs for cross-dataset generalisation (train OFR, test ONR) | FCN-8s-FL · ONR | F-measure (%): 9.16% -> 31.43% | Improved |
| add RAUs for cross-dataset generalisation (train ONR, test OFR) | FCN-8s-FL · OFR | F-measure (%): 22.98% -> 36.60% | Improved |
| add RAUs to FCN-8s-FL (inference time cost) | FCN-8s-FL · ONR | Inference time per frame (s): 0.06s -> 0.32s | Hurt |
Closest papers by meaning
- Joint Rain Detection and Removal from a Single Image with Contextualized Deep Networks (TPAMI 2019 · similarity 0.85)
- Looking Into the Water by Unsupervised Learning of the Surface Shape (NeurIPS 2025 · similarity 0.85)
- Removing Rain From Single Images via a Deep Detail Network (CVPR 2017 · similarity 0.84)
- DSNet: Joint Semantic Learning for Object Detection in Inclement Weather Conditions (TPAMI 2020 · similarity 0.84)
- R-FCN++: Towards Accurate Region-Based Fully Convolutional Networks for Object Detection (AAAI 2018 · similarity 0.84)
- GLH-Water: A Large-Scale Dataset for Global Surface Water Detection in Large-Size Very-High-Resolution Satellite Imagery (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.
