Published at ICCV 2025: 'Perceiving and Acting in First-Person: A Dataset and Benchmark for Egocentric Human-Object-Human Interactions' – introduced InterVLA, the first large-scale egocentric human-object-human interaction dataset
Published at ECCV 2024: 'HIMO: A New Benchmark for Full-Body Human Interacting with Multiple Objects' – built a large-scale dataset with 3.3K HOI sequences and 4.08M frames
Published at CVPR 2024: 'Inter-X: Towards Versatile Human-Human Interaction Analysis' – proposed a large-scale human-human interaction dataset with 11K sequences and over 8.1M frames
Published at CVPR 2024: 'ReGenNet: Towards Human Action-Reaction Synthesis' – introduced the first multi-setting benchmark for human action-reaction synthesis
Published at ICCV 2023: 'ActFormer: A GAN-based Transformer towards General Action-Conditioned 3D Human Motion Generation' – presented a GAN-based Transformer for action-conditioned 3D motion generation
Co-authored arXiv 2025 paper 'TeleOpBench: A Simulator-Centric Benchmark for Dual-Arm Dexterous Teleoperation' – developed a modular benchmark for dual-arm teleoperation
Published arXiv 2024: 'MotionBank: A Large-scale Video Motion Benchmark with Disentangled Rule-based Annotations' – proposed a large video motion benchmark with rule-based disentangled annotations
Published in TMM 2022: 'Skeleton-Based Mutually Assisted Interacted Object Localization and Human Action Recognition' – proposed a joint learning framework for object localization and action recognition