Namhoon Jung

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

Yechan Kim, JongHyun Park, Dongho Yoon, Namhoon Jung, Moongu Jeon

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 alignme…Read more