UniHuman: A Unified Model For Editing Human Images in the Wild

In the following, we first discuss related literature in Sec. 1. Then we analyze our text manipulation results in Sec. 2 and show more ablation experiments in Sec. 3. Subsequently, we explain how to incorporate LH-400K dataset into training in Sec. 4 and show that our unified model can achieve multi-task combinations in Sec. 5. Finally, we explain our implementation details in Sec. 6 and present various visualized examples in Sec. 8.