issues about mano param in freihand dataset

mks0601 / I2L-MeshNet_RELEASE

Official PyTorch implementation of "I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Image", ECCV 2020

MIT License

724 stars 127 forks source link

issues about mano param in freihand dataset #119

Open vincentkeino opened 1 year ago

vincentkeino commented 1 year ago

hi,i just use your code for estimate hands,but i have some problem of mano params.In dataset of freihand you provided,1-48 of mano params are pose of wrist and hands,49-58 are shape params ,59-61 is translate of hands. did i right? The trans used in your dataset has been changed to cam-coordination.did i undersrand right ? please help affirm it!

mks0601 commented 1 year ago

Please read README in https://drive.google.com/drive/folders/13qR8EhHFgvJ_AjgHz-JZKQQzYJX7oLpE