analiza aiqbench
HDST-GNN addresses critical limitations in UAV multi-object tracking by introducing altitude-adaptive graph connectivity and heterogeneous node representations to handle occlusions and varying flight conditions. The model significantly reduces identity switches compared to standard trackers like SORT, demonstrating robust performance even with imperfect detection inputs from YOLOv8n. This work offers a practical advancement for computer vision systems operating in complex aerial environments where object scale and visibility fluctuate dynamically.
HDST-GNN: Heterogeneous Dynamic Spatiotemporal Graph Neural Networks for Multi-Object Tracking in UAV Aerial Imagery
arXiv:2606.05587v1 Announce Type: cross Abstract: Multi-object tracking (MOT) from UAV imagery presents unique challenges: altitude varies across sequences, objects are small and densely packed, and frequent occlusion causes identity switches. Existing graph-based trackers assume fixed spatial context and treat all objects uniformly, ignoring the heterogeneous lifecycle states of detections, active tracklets, and lost targets. We propose HDST-GNN, a Heterogeneous Dynamic Spatiotemporal Graph Neural Network with three novel contributions. First, Altitude-Adaptive Edge Construction estimates a camera-altitude proxy from mean object area and adjusts the graph connectivity radius accordingly. Second, Heterogeneous Node Representation models detections (Type-D), confirmed tracklets (Type-T), and lost tracklets (Type-L) as distinct node types with dedicated projections and typed edge relations. Third, Occlusion-Gated Temporal Aggregation gates each node's attention contribution by its occlusion confidence, preventing occluded nodes from corrupting neighbour embeddings. HDST-GNN is trained end-to-end with a differentiable Sinkhorn head using joint cross-entropy and triplet loss. On VisDrone2019-MOT with oracle detections, HDST-GNN achieves 94.51% MOTA and 97.24% IDF1, outperforming SORT by +5.0 MOTA points and reducing identity switches by 81%. With real YOLOv8n detections, HDST-GNN reduces identity switches by 49% vs. SORT. Ablation studies confirm the independent contribution of each component.
Abstract:Multi-object tracking (MOT) from UAV imagery presents unique challenges: altitude varies across sequences, objects are small and densely packed, and frequent occlusion causes identity switches. Existing graph-based trackers assume fixed spatial context and treat all objects uniformly, ignoring the heterogeneous lifecycle states of detections, active tracklets, and lost targets. We propose HDST-GNN, a Heterogeneous Dynamic Spatiotemporal Graph Neural Network with three novel contributions. First, Altitude-Adaptive Edge Construction estimates a camera-altitude proxy from mean object area and adjusts the graph connectivity radius accordingly. Second, Heterogeneous Node Representation models detections (Type-D), confirmed tracklets (Type-T), and lost tracklets (Type-L) as distinct node types with dedicated projections and typed edge relations. Third, Occlusion-Gated Temporal Aggregation gates each node's attention contribution by its occlusion confidence, preventing occluded nodes from corrupting neighbour embeddings. HDST-GNN is trained end-to-end with a differentiable Sinkhorn head using joint cross-entropy and triplet loss. On VisDrone2019-MOT with oracle detections, HDST-GNN achieves 94.51% MOTA and 97.24% IDF1, outperforming SORT by +5.0 MOTA points and reducing identity switches by 81%. With real YOLOv8n detections, HDST-GNN reduces identity switches by 49% vs. SORT. Ablation studies confirm the independent contribution of each component.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2606.05587 [cs.CV]
(or arXiv:2606.05587v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2606.05587 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 290 cuvinte
Subiecte conexe
PPO-STGNN: A Proximal Policy Optimization Approach with Spatio-Temporal Graph Neural Networks for DAG Task Scheduling in Cloud-Edge-End Computing
arXiv · 5 sept. 2026
K-STEMIT: Knowledge-Informed Spatio-Temporal Efficient Multi-Branch Graph Neural Network for Subsurface Stratigraphy Thickness Estimation from Radar Data
arXiv · 4 aug. 2026
ChronoSpike: An Adaptive Spiking Graph Neural Network for Dynamic Graphs
arXiv · 5 oct. 2026
Coordinating from Memory: Graph-Structured Experience Reuse for Multi-Agent Adaptation in Dynamic Manufacturing
arXiv · 23 iul. 2026