再读MANO / SMPL build parametric model 部分

描述数据集的内容和分布

The multi-pose dataset consists of 1786 registrations of 40 individuals (891 registrations spanning 20 females, and 895 registrations spanning 20 males);

spanning: a range or variety of sth

====
We alternate between optimizing registration specific parameters \theta, subject-specific parameters {\TˆP_i , \JˆP_i }, and global parameters {W, P}.

alternate between: 交替/轮流

====
To help prevent overfitting of the pose-dependent blend shapes, we regularize them towards zero

to help prevent ...

====
To that end: 所以

====
We tried several regression strategies; what we found to work best, was to compute J using non-negative least squares [Lawson and Hanson 1995] with the inclusion of a term that encourages the weights to add to one.