Realistically distributing object placements in synthetic training data improves the performance of vision-based object detection models.
Setareh DabiriVasileios LioutasBerend ZwartsenbergYunpeng LiuMatthew NiedobaXiaoxuan LiangDylan GreenJustice SefasJonathan Wilder LavingtonFrank WoodAdam ScibiorPublished in: CoRR (2023)