Research map / Label-efficient and robust learning / Recognition and long-tail
RareSpot+: A Benchmark, Model, and Active Learning Framework for Small and Rare Wildlife in Aerial Imagery
Bowen Zhang, Jesse T. Boulerice, Charvi Mendiratta, Nikhil Kuniyil, Satish Kumar, Hila Shamon, Bangalore S Manjunath
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 Multi-Scale Consistency Learning (ablation) | YOLOv5L · Prairie dog benchmark (test) | overall mAP@50: 0.626 -> 0.684 | Improved |
| add Context-Aware Hard Sample Augmentation (ablation) | YOLOv5L · Prairie dog benchmark (test) | overall mAP@50: 0.626 -> 0.647 | Improved |
| combine MSCL and CA-HSA (RareSpot+) (ablation) | YOLOv5L · Prairie dog benchmark (val) | overall mAP@50: 0.620 -> 0.681 | Improved |
| combine MSCL and CA-HSA (RareSpot+) (ablation) | YOLOv5L · Prairie dog benchmark (test) | prairie dog AP@50: 0.366 -> 0.495 | Improved |
| add CA-HSA to YOLOv5L (ablation) | YOLOv5L · Prairie dog benchmark (test) | prairie dog precision: 0.482 -> 0.464 | Hurt |
| replace YOLOv5L with newer or transformer detectors (ablation) | YOLOv7L to Co-DETR · Prairie dog benchmark (val) | overall mAP@50: 0.620 -> 0.517 (YOLOv7L), 0.411 (DETR), 0.437 (Co-DETR) | Hurt |
| restrict YOLOv5L to P3 head only (ablation) | YOLOv5L · Prairie dog benchmark (val) | overall mAP@50: 0.620 -> 0.623 | No clear effect |
| add MSCL to YOLOv5 (ablation) | YOLOv5 · WAID | mAP@50: 0.956 -> 0.976 | Improved |
| add MSCL to YOLOv5 (ablation) | YOLOv5 · AED | mAP@50: 0.891 -> 0.897 | No clear effect |
| add MSCL and CA-HSA to YOLOv5L (ablation) | YOLOv5L · HerdNet | mAP@50: 0.773 -> 0.870 | Improved |
| add MSCL and CA-HSA to YOLOv5L (ablation) | YOLOv5L · Eikelboom | mAP@50: 0.830 -> 0.852 | Improved |
| add geospatial prior to random sampling for active learning | RareSpot+ active learning · Prairie dog active learning | PD recall: 0.539 -> 0.568 | Improved |
Closest papers by meaning
- WildlifeMapper: Aerial Image Analysis for Multi-Species Detection and Identification (CVPR 2024 · similarity 0.89)
- BirdCollect: A Comprehensive Benchmark for Analyzing Dense Bird Flock Attributes (AAAI 2024 · similarity 0.88)
- BuckTales: A multi-UAV dataset for multi-object tracking and re-identification of wild antelopes (NeurIPS 2024 · similarity 0.87)
- WildBox: A Dataset and Benchmark for Aerial Monocular 3D Detection of African Savanna Wildlife (NeurIPS 2026 · similarity 0.87)
- MammAlps: A Multi-view Video Behavior Monitoring Dataset of Wild Mammals in the Swiss Alps (CVPR 2025 · similarity 0.87)
- Reconstructing Animals and the Wild (CVPR 2025 · similarity 0.85)
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.
