Abstract: This paper presents a method to learn hand-object interaction prior for reconstructing a 3D hand-object scene from a single RGB image. The inference as well as training-data generation for ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results