MVA 2025 Small Multi-Object Tracking for Spotting Birds Challenge: Dataset, Methods, and Results

2025/07/17 by Yuki Kondo, Kondo, Yuki, Norimichi Ukita +43 · 1 citation
Agricultural and Biological Sciences · #Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Smart Agriculture and AI

paper · pdf · doi:10.48550/arxiv.2507.12832

openalex publication_date 2025/07/17 · openalex created_date 2025/10/18 · openalex updated_date 2026/07/28

Abstract

Small Multi-Object Tracking (SMOT) is particularly challenging when targets occupy only a few dozen pixels, rendering detection and appearance-based association unreliable. Building on the success of the MVA2023 SOD4SB challenge, this paper introduces the SMOT4SB challenge, which leverages temporal information to address limitations of single-frame detection. Our three main contributions are: (1) the SMOT4SB dataset, consisting of 211 UAV video sequences with 108,192 annotated frames under diverse real-world conditions, designed to capture motion entanglement where both camera and targets move freely in 3D; (2) SO-HOTA, a novel metric combining Dot Distance with HOTA to mitigate the sensitivity of IoU-based metrics to small displacements; and (3) a competitive MVA2025 challenge with 78 participants and 308 submissions, where the winning method achieved a 5.1x improvement over the baseline. This work lays a foundation for advancing SMOT in UAV scenarios with applications in bird strike avoidance, agriculture, fisheries, and ecological monitoring.

Citations

• GPLQ: A General, Practical, and Lightning QAT Method for Vision Transformers
• QwT-v2: Practical, Effective and Efficient Post-Training Quantization
• YOLOv12: Attention-Centric Real-Time Object Detectors
• Quantization without Tears
• FOLT: Fast Multiple Object Tracking from UAV-captured Videos Based on Optical Flow
• Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking
• BandRe: Rethinking Band-Pass Filters for Scale-Wise Object Detection Evaluation
• LightGlue: Local Feature Matching at Light Speed
• Iterative Scale-Up ExpansionIoU and Deep Features Association for Multi-Object Tracking in Sports
• AirBirds: A Large-scale Challenging Dataset for Bird Strike Prevention in Real-world Airports
• DETRs with Collaborative Hybrid Assignments Training
• RFLA: Gaussian Receptive Field based Label Assignment for Tiny Object Detection
• YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors
• YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors
• BoT-SORT: Robust Associations Multi-Pedestrian Tracking
• Observation-Centric SORT: Rethinking SORT for Robust Multi-Object Tracking
• Tiny Object Tracking: A Large-scale Dataset and A Baseline
• High-Resolution Image Synthesis with Latent Diffusion Models
• High-Resolution Image Synthesis with Latent Diffusion Models
• ByteTrack: Multi-Object Tracking by Associating Every Detection Box
• ByteTrack: Multi-object Tracking by Associating Every Detection Box
• YOLOX: Exceeding YOLO Series in 2021
• MOTR: End-to-End Multiple-Object Tracking with Transformer
• USB: Universal-Scale Object Detection Benchmark
• TrackFormer: Multi-Object Tracking with Transformers
• HOTA: A Higher Order Metric for Evaluating Multi-Object Tracking
• Rethinking Channel Dimensions for Efficient Model Design
• DISK: Learning local features with policy gradient
• COMET: Context-Aware IoU-Guided Network for Small Object Tracking
• FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Tracking
• MOT20: A benchmark for multi object tracking in crowded scenes
• Detection and Tracking Meet Drones Challenge
• Scale Match for Tiny Person Detection
• Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression
• Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression
• Towards Real-Time Multi-Object Tracking
• Objects as Points
• Task-Driven Super Resolution: Object Detection in Low-resolution Images
• The Unmanned Aerial Vehicle Benchmark: Object Detection and Tracking
• Cascade R-CNN: Delving into High Quality Object Detection
• Feature Pyramid Networks for Object Detection
• Performance Measures and a Data Set for Multi-Target, Multi-Camera Tracking
• MOT16: A Benchmark for Multi-Object Tracking
• MOTChallenge 2015: Towards a Benchmark for Multi-Target Tracking
• Microsoft COCO: Common Objects in Context
• Social Force Model for Pedestrian Dynamics
• Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
• Swin Transformer: Hierarchical Vision Transformer using Shifted Windows

Cited by

• OmniTrack++: Omnidirectional Multi-Object Tracking by Learning Large-FoV Trajectory Feedback
• YOLOv8-SMOT: An Efficient and Robust Framework for Real-Time Small Object Tracking via Slice-Assisted Training and Adaptive Association

Related

• YOLOv8-SMOT: An Efficient and Robust Framework for Real-Time Small Object Tracking via Slice-Assisted Training and Adaptive Association
• Improving Small Drone Detection Through Multi-Scale Processing and Data Augmentation
• DroneMOT: Drone-based Multi-Object Tracking Considering Detection Difficulties and Simultaneous Moving of Drones and Objects
• Small Bird Detection using YOLOv7 with Test-Time Augmentation
• Multi-Drone based Single Object Tracking with Agent Sharing Network