G-MAD: A Game-Based Data Generation Framework for Multi-View RGB-T Aerial Object Detection

2026-07-22

AI Summary

This work addresses key challenges in real-world aerial datasets—limited viewpoints, poor alignment between RGB and thermal (RGB-T) modalities, and high annotation costs—by introducing G-MAD, the first open-source framework built upon the Arma3 game engine. G-MAD enables structured scene configuration, controllable multi-view camera placement, and synchronized acquisition of geometrically aligned RGB-T image pairs. Leveraging the engine’s intrinsic geometric metadata, it further supports automatic bounding box annotation. Using this framework, the authors construct and release AMOD, a large-scale, multi-view aerial RGB-T object detection benchmark dataset. This contribution substantially reduces the cost of acquiring and annotating high-quality multimodal aerial data, offering a controllable experimental platform for research on viewpoint variation, multimodal fusion, and synthetic-to-real domain transfer.

Abstract

This work introduces G-MAD, an open-source framework that uses Arma3 to generate synchronized multi-view RGB-T data for aerial object detection. G-MAD addresses key limitations of real-world aerial dataset construction, including limited viewpoint control, imperfect RGB-T alignment and high annotation cost. The framework supports structured scenario specification, controllable multi-view camera placement, simultaneous visible/thermal capture, and automatic bounding box annotation using engine-level geometric metadata. These capabilities enable controlled studies of viewpoint variation, multi-modal fusion, and synthetic-to-real transfer in aerial object detection. Besides, using G-MAD, we construct and release AMOD, a new large-scale multi-view aerial RGB-T object detection benchmark. The source code and the dataset are available at https://unique-chan.github.io/G-MAD-Project.

Problem

• aerial object detection
• RGB-T alignment
• viewpoint control
• annotation cost
• multi-view data

Innovation

• game-based simulation
• multi-view RGB-T
• aerial object detection
• synthetic data generation
• automatic annotation