Skip to content

Latest commit

 

History

History

preprocess

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
frame_extraction.py
frame_extraction.py
 
 
to_sequence_data.py
to_sequence_data.py
 
 

README.md

Data Pre-process

  1. Download the AMASS dataset. Specifically, in this project, we use the following subsets
['Transitions_mocap', 'BMLmovi', 'MPI_mosh', 'MPI_Limits', 'SSM_synced', 'SFU', 'MPI_HDM05', 'TCD_handMocap', 
'HumanEva', 'TotalCapture', 'DFaust_67', 'ACCAD', 'CMU', 'EKUT', 'Eyes_Japan_Dataset', 'KIT']
  1. Put the downloaded data in directory raw_data as follows:
raw_data
├── Transitions_mocap
├── BMLmovi
├── MPI_mosh
| ...
  1. Extract each frame from the data with the rate of 30 fps
python frame_extraction.py
  1. Save the necessary parameters and merge into 128 frames per sequence.
python to_sequence_data.py

The generated data file is amass_smpl_30fps_128frame.npz