Research map / Detection and segmentation / Open-set and few-shot detection
Urban Waterlogging Detection: A Challenging Benchmark and Large-Small Model Co-Adapter
Suqi Song, Chenxu Zhang, Peng Zhang, Pengkun Li, Fenglong Song, Lei Zhang
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 |
|---|---|---|---|
| use two-stage instead of one-stage joint training of large and small models (ablation) | LSM-Adapter · UW-Bench UW-all | IoU: 44.86 -> 57.73 | Improved |
| add histogram equalization adapter to SAM-Adapter (ablation) | SAM-Adapter · UW-Bench UW-all | IoU: 51.16 -> 55.83 | Improved |
| add spatial prompt (ablation) | LSM-Adapter_S with HE-Adapt · UW-Bench UW-all | IoU: 55.83 -> 58.67 | Improved |
| add semantic prompt (ablation) | LSM-Adapter_S · UW-Bench UW-all | IoU: 58.67 -> 58.71 | No clear effect |
| add style prompt (ablation) | LSM-Adapter_U · UW-Bench UW-all | IoU: 56.45 -> 57.14 | Improved |
| add dynamic prompt combiner (ablation) | LSM-Adapter_U · UW-Bench UW-all | IoU: 57.14 -> 58.42 | Improved |
| use mask prompt rather than point prompt (ablation) | LSM-Adapter_M two-stage · UW-Bench UW-all | IoU: 55.85 -> 57.73 | Improved |
| use mask rather than box prompt (ablation) | LSM-Adapter_U two-stage · UW-Bench UW-all | IoU: 58.45 (box) vs 58.42 (mask) | No clear effect |
| co-train small model with SAM (LSM-Adapter) vs small model alone | Mask R-CNN · UW-Bench UW-all | IoU: 37.84 -> 57.73 | Improved |
| LSM-Adapter vs Segformer on precision | UW-all · UW-Bench UW-all | Precision: 86.63 (Segformer) vs 79.47 (LSM-Adapter_S) | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Segformer — SegFormer: A Topic Segmentation Model with Controllable Range of Attention (AAAI 2023) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- GLH-Water: A Large-Scale Dataset for Global Surface Water Detection in Large-Size Very-High-Resolution Satellite Imagery (AAAI 2024 · similarity 0.83)
- UrbanDataLayer: A Unified Data Pipeline for Urban Science (NeurIPS 2024 · similarity 0.82)
- LAqua: Laplacian Pyramids for Aquatic Segmentation (Student Abstract) (AAAI 2025 · similarity 0.81)
- Diving into Underwater: Segment Anything Model Guided Underwater Salient Instance Segmentation and A Large-scale Dataset (ICML 2024 · similarity 0.81)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.81)
- Semantic Segmentation on Raindrop Degraded Images Using Two-Stage Dual Teacher-Student Learning (AAAI 2025 · similarity 0.80)
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.
