5 papers
From Draft to Draft-Free: One-Step Video Object Removal via Privileged Distillation and Fast Planting
Zizhao Chen, Ping Wei, Guang Dai +2
Video object removal is a fundamental yet challenging task in video editing. Despite recent progress, existing methods typically fall into two categories. Traditional approaches ba…
Differential Amplifier-Inspired AmpAttention for Multi-View Robotic Manipulation
Jin Yang, Ping Wei, Nanning Zheng
Multi-view robotic manipulation methods with the attention mechanism have recently achieved significant progress in both training efficiency and task performance. However, the inhe…
Hyperbolic Multiview Pretraining for Robotic Manipulation
Jin Yang, Ping Wei, Yixin Chen +1
3D-aware visual pretraining has proven effective in improving the performance of downstream robotic manipulation tasks. However, existing methods are constrained to Euclidean embed…
-World: Intra-Inter Tokenization for Efficient Dynamic 4D Scene Forecasting
Zhimin Liao, Ping Wei, Ruijie Zhang +3
Forecasting the evolution of 3D scenes and generating unseen scenarios via occupancy-based world models offers substantial potential for addressing corner cases in autonomous drivi…
Task-Driven Exploration: Decoupling and Inter-Task Feedback for Joint Moment Retrieval and Highlight Detection
Jin Yang, Ping Wei, Huan Li +1
Video moment retrieval and highlight detection are two highly valuable tasks in video understanding, but until recently they have been jointly studied. Although existing studies ha…