Research map / Detection and segmentation / Tracking
Detection, Tracking, and Counting Meets Drones in Crowds: A Benchmark
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 STNNet instead of DM-Count for density estimation | DroneCrowd · DroneCrowd test | MAE: 2.6 MAE | Helped |
| Use STNNet instead of DM-Count for density estimation | DroneCrowd · DroneCrowd test | MSE: 8.3 MSE | Helped |
| Use the STNNet localization approach instead of DM-Count localization | DroneCrowd · DroneCrowd test | L-mAP: 40.45% vs. 18.17% | Helped |
| Use STNNet with min-cost flow tracking instead of DM-Count with min-cost flow | DroneCrowd · DroneCrowd test | T-mAP: 32.50% vs. 17.01% | Helped |
| Add the localization subnet (ablation) | STNNet · DroneCrowd test | MAE: 18.6 -> 15.8 | Helped |
| Add the localization subnet (ablation) | STNNet · DroneCrowd test | MSE: 22.2 -> 18.7 | Helped |
| Add the localization subnet (ablation) | STNNet · DroneCrowd test | L-mAP: 32.19% -> 40.45% | Helped |
| Add the association subnet (ablation) | STNNet · DroneCrowd test | L-mAP: 39.77% -> 40.45% | Helped |
| Add predicted motion-offset association (ablation) | STNNet · DroneCrowd test | T-mAP: 31.44% vs. 32.32% | Helped |
| Add the relation constraint to the neighboring context loss (ablation) | STNNet · DroneCrowd test | L-mAP: 40.00% -> 40.23% | Helped |
| Add the relation constraint to the neighboring context loss (ablation) | STNNet · DroneCrowd test | T-mAP: 32.26% -> 32.50% | Helped |
| Use cycle-based forward and backward offsets instead of forward offsets only (ablation) | STNNet · DroneCrowd test | L-mAP: 40.23% -> 40.45% | Helped |
Closest papers by meaning
- Adapting Lightweight Image-based Counting Models for Video Crowd Counting (CVPR 2026 · similarity 0.86)
- Dogfight: Detecting Drones From Drones Videos (CVPR 2021 · similarity 0.84)
- PointSR: Self-Regularized Point Supervision for Drone-View Object Detection (CVPR 2025 · similarity 0.83)
- RemDet: Rethinking Efficient Model Design for UAV Object Detection (AAAI 2025 · similarity 0.82)
- Weakly Supervised Video Individual Counting (CVPR 2024 · similarity 0.82)
- Point-to-Region Loss for Semi-Supervised Point-Based Crowd Counting (CVPR 2025 · similarity 0.82)
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.