Hengjia Li

6

Papers

18

Total Citations

Papers (6)

Local Conditional Controlling for Text-to-Image Diffusion Models

Pseudo Label Refinery for Unsupervised Domain Adaptation on Cross-dataset 3D Object Detection

PersonalVideo: High ID-Fidelity Video Customization without Dynamic and Semantic Degradation

MagicID: Hybrid Preference Optimization for ID-Consistent and Dynamic-Preserved Video Customization

DreamRelation: Relation-Centric Video Customization

TagCLIP: A Local-to-Global Framework to Enhance Open-Vocabulary Multi-Label Classification of CLIP without Training