Tao Hu

InterSketch shows that interleaving visual sketches with textual reasoning, guided by self-correction and stepwise rewards, unlocks surprisingly strong long-horizon visual reasoning, even surpassing Gemini-3-Pro.

Ditch the Gaussian noise: HumANDiff generates more realistic human videos by sampling noise directly on a 3D articulated human body, ensuring motion consistency.

Generate realistic, controllable human videos from a single image with 3D pose and view control, finally solving the wrinkle-rendering problem.