IEEE transactions on visualization and computer graphics
SeparateGen: Semantic Component-based 3D Character Generation from Single Images.
Dong-Yang Li, Yi-Long Liu, Zi-Xian Liu, Yan-Pei Cao, Meng-Hao Guo, Shi-Min Hu
Published: 202610.1109/TVCG.2026.3652452
Abstract
Creating detailed 3D characters from a single image remains challenging due to the difficulty in separating semantic components during generation. Existing methods often produce entangled meshes with poor topology, hindering downstream applications l…
Preview only. Read the full abstract at the source