no code implementations • 26 Dec 2023 • Hyun Kang, Dohae Lee, Myungjin Shin, In-Kwon Lee
Based on this observation, we devise a technique to steer the guidance direction towards any concept from which the model diverges.
no code implementations • 30 Aug 2023 • Myungjin Shin, Dohae Lee, In-Kwon Lee
This is based on the premise that the ultimate goal of pose reconstruction is to reconstruct the motion, which is a series of poses.
no code implementations • 7 Apr 2023 • Dohae Lee, Hyun Kang, In-Kwon Lee
We present ClothCombo, a pipeline to drape arbitrary combinations of clothes on 3D human models with varying body shapes and poses.
1 code implementation • 7 Dec 2021 • Jinhyung Park, Dohae Lee, In-Kwon Lee
This precondition restrains the model from generalizing to real-world data, which is considered to be a sequence of unordered point sets.
no code implementations • 10 Nov 2021 • Dohae Lee, Young Jin Oh, In-Kwon Lee
However, existing approaches are difficult to apply if the number of relations is unknown or if the relations are complex.