Research map / Video / Video segmentation and tracking
Video Individual Counting for Moving Drones
Yaowu Fan, Jia Wan, Tao Han, Antoni B. Chan, Andy J. Ma
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 |
|---|---|---|---|
| estimate shared density maps first and derive inflow by subtraction, instead of learning inflow directly (ablation) | SDNet · MovingDroneCrowd test | MAE: 65.64 -> 41.00 | Improved |
| use depth-wise multi-scale cross-frame attention (DCFA) instead of shallow attention on global features (ablation) | SDNet · MovingDroneCrowd test | MAE: 70.42 -> 41.00 | Improved |
| add positional encoding in DCFA with VGG backbone (ablation) | SDNet (VGG16) · MovingDroneCrowd test | MAE: 41.00 -> 66.64 | Hurt |
| add positional encoding in DCFA with ViT backbone (ablation) | SDNet (ViT) · MovingDroneCrowd test | MAE: 98.56 -> 51.76 | Improved |
| use VGG16 CNN backbone instead of ViT (ablation) | SDNet · MovingDroneCrowd test | MAE: 51.76 -> 41.00 | Improved |
| use shared-density-map approach versus localization-then-association VIC (CGNet) | SDNet · MovingDroneCrowd test | MAE: 66.06 -> 41.00 | Improved |
| use SDNet versus MOT trackers on sparse density level D1 | SDNet · MovingDroneCrowd test (D1) | MAE: 24.00 (ByteTrack) vs 79.77 (ours) | Hurt |
| use SDNet versus CGNet on dynamic scenes of UAVVIC | SDNet · UAVVIC val (dynamic) | MSE: 30.62 -> 7.33 | Improved |
Closest papers by meaning
- Detection, Tracking, and Counting Meets Drones in Crowds: A Benchmark (CVPR 2021 · similarity 0.89)
- Weakly Supervised Video Individual Counting (CVPR 2024 · similarity 0.88)
- Drone-Based Object Counting by Spatially Regularized Regional Proposal Network (ICCV 2017 · similarity 0.87)
- DR.VIC: Decomposition and Reasoning for Video Individual Counting (CVPR 2022 · similarity 0.87)
- Adapting Lightweight Image-based Counting Models for Video Crowd Counting (CVPR 2026 · similarity 0.86)
- Efficient Masked AutoEncoder for Video Object Counting and A Large-Scale Benchmark (ICLR 2025 · similarity 0.86)
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.
