VETD220

Official repository of VETD220, a visible-event multimodal benchmark for anti-UAV tracking.

Overview

Anti-UAV tracking in real-world environments is challenging due to factors such as fast motion, small targets, illumination variation, background clutter, and motion blur. To support research in this direction, we present VETD220, a visible-event multimodal dataset designed for anti-UAV tracking.

VETD220 contains 220 videos and 68,379 visible-event sequence pairs, covering 6 real-world scenarios and 15 challenging attributes. The dataset provides synchronized visible frames and event streams, supporting both model training and fair evaluation for visible-event anti-UAV tracking.

Features

• Visible-event multimodal data for anti-UAV tracking
• 220 videos and 68,379 visible-event sequence pairs
• 6 real-world scenarios
• 15 challenging attributes
• Supports both training and evaluation
• Designed for single-object tracking

Benchmark Task

VETD220 is designed for single-object anti-UAV tracking with visible-event multimodal input.

Evaluation Metrics

We adopt the standard single-object tracking metrics:

• SR: Success Rate
• PR: Precision Rate

Dataset Split

The dataset is divided into:

• Training set: 160 videos
• Validation set: 20 videos
• Test set: 40 videos

Dataset Structure

The expected folder structure is as follows:

VETD220/
├── train/
│   ├── sequence_001/
│   │   ├── aps/
│   │   ├── event/
│   │   └── label.txt
│   ├── sequence_002/
│   │   ├── aps/
│   │   ├── event/
│   │   └── label.txt
│   ├──  ...
│   └──  list_train.txt
├── test/
│   ├── sequence_001/
│   │   ├── aps/
│   │   ├── event/
│   │   └── label.txt
│   ├── sequence_002/
│   │   ├── aps/
│   │   ├── event/
│   │   └── label.txt
│   ├──  ...
│   └──  list_test.txt

The complete dataset and related resources will be made publicly available soon.