G-MAD is an open-source framework built on Arma 3 for generating synchronized multi-view RGB-T data for aerial object detection. It provides structured scenario specification, controllable camera placement, simultaneous visible and thermal capture, and automatic bounding-box annotation from engine-level geometric metadata. The authors also introduce AMOD, a large-scale multi-view aerial RGB-T benchmark. The framework is intended to support controlled studies of viewpoint variation, multimodal fusion, and synthetic-to-real transfer while reducing the cost and alignment problems associated with real-world aerial dataset construction. The supplied abstract does not report benchmark results or transfer performance.
No heat snapshots are available in the last 24 hours.