Honda/UCSD

The dataset contains 59 videos of 20 subjects with varying poses and expressions.

BibTex: