HumanDiT: Pose-Guided Diffusion Transformer for Long-form Human Motion Video Generation

Qijun Gan, Yi Ren, Chen Zhang, Zhenhui Ye, Pan Xie, Xiang Yin, Zehuan Yuan, BINGYUE PENG, Jianke Zhu

Project

• HumanDiT is a new pose-guided Diffusion Transformer (DiT) framework for human motion video generation. It addresses challenges in fine-grained body rendering, resolution flexibility, and visual consistency by leveraging a large-scale dataset and advanced architectural design.

PianoMotion10M: Dataset and Benchmark for Hand Motion Generation in Piano Performance

Qijun Gan, Song Wang, Shengtao Wu, Jianke Zhu

ICLR 2025 (Spotlight - Top 5%)

Project

• PianoMotion10M is a new benchmark for piano-hand motion generation, providing guidance on hand movements and fingerings in piano playing.

XHand: Real-time Expressive Hand Avatar

Qijun Gan, Zijie Zhou, Jianke Zhu

Project

• XHand is a real-time expressive hand avatar designed for high-fidelity hand shape, appearance, and deformation generation.

Fine-grained multi-view hand reconstruction using inverse rendering

Qijun Gan, Wentong Li, Jinwei Ren, Jianke Zhu

Project

• We propose a fine-grained multi-view hand mesh reconstruction method that leverages inverse rendering to recover detailed hand poses and textures.

FastHuman: Reconstructing High-Quality Clothed Human in Minutes

Lixiang Lin, Songyou Peng, Qijun Gan, Jianke Zhu

Project

• We introduce a fast optimization approach for high-quality clothed human body shape reconstruction from multi-view posed images.