nkolot / GraphCMR

Repository for the paper "Convolutional Mesh Regression for Single-Image Human Shape Reconstruction"
BSD 3-Clause "New" or "Revised" License
425 stars 67 forks source link

Generate 3D mesh for Human36M #9

Closed jeroenvb closed 5 years ago

jeroenvb commented 5 years ago

Very nice work. The other thread is closed, but I'd like to understand how we may generate the 3D (smpl) meshes for the Human36M dataset.

You mention on your github that you MoSHed the data to get SMPL parameters. On the mosh project page we don't see any code that performs this.

We are training a modified network from scratch, and need more ground truth 3D meshes. Any info would be much appreciated!

geopavlakos commented 5 years ago

For Human3.6M, the SMPL parameters we used for training are the same as the ones used by the HMR approach (https://github.com/akanazawa/hmr). Unfortunately, this data is no longer publicly available due to license limitations. Beyond that, for this work we haven't investigated generating ground truth 3D meshes for Human3.6M.