Hand Motion Generation

RoMo Hands: A Large Scale Richly Organized Text to Hand Motion Dataset
RoMo Hands is a large-scale text-paired hand motion dataset with approximately 542K bimanual sequences, five caption granularities, and a unified 393-D representation for evaluating text-to-hand motion generation.
Yizhak Ben-Shabat, Jiahao Zhang, Joseph Liu, Seonghyeon Moon, Young-Yoon Lee, Haomiao Jiang, Oren Jacob, Mubbasir Kapadia
PDF Cite Dataset Project
RoMo Hands: A Large Scale Richly Organized Text to Hand Motion Dataset