AI RESEARCH
THFM: A Unified Video Foundation Model for 4D Human Perception and Beyond
arXiv CS.CV
•
ArXi:2603.25892v1 Announce Type: new We present THFM, a unified video foundation model for human-centric perception that jointly addresses dense tasks (depth, normals, segmentation, dense pose) and sparse tasks (2d/3d keypoint estimation) within a single architecture. THFM is derived from a pretrained text-to-video diffusion model, repurposed as a single-forward-pass perception model and augmented with learnable tokens for sparse predictions. Modulated by the text prompt, our single unified model is capable of performing various perception tasks.