Skip to content

Latest commit

 

History

History
33 lines (24 loc) · 2.16 KB

ReadMe.md

File metadata and controls

33 lines (24 loc) · 2.16 KB

How to Generate SewFactory Dataset

Installation (Windows, GPU Enabled)

  1. Maya python environment:

  • Dependencies: Autodesk Maya 2020 with two plugins: Arnold and Qualoth

    We have only tested the sewfactory on this version of dependencies, and using the latest version may result in errors. Please prepare Arnold license to render images without watermarks.

  • Maya Python Environment: Numpy for python2.7

  1. Windows python environment:
  • python3.9 -m pip install numpy scipy svglib svgwrite psutil wmi

Configuration

  1. cd path_to_dev/SewFactory and add path_to_dev/SewFactory/packages to PYTHONPATH on your system for correct importing of our custom modules.

  2. Download SMPLH Fbx model (SMPLH_female_010_207.fbx) to meta_infos\\fbx_metas.

  3. (Optional) Prepare human skin textures to examples\\skin_textures (for example, SURREAL).

  4. Prepare human poses to examples\\human_poses (for example, AMASS)

  5. Collect some images used for garment textures and put them into examples\\garment_textures

Generation

  1. Sample sewing pattern designs from the template (Please update the parameters in meta_infos\\configs\\dataset_config.yaml to generate your dataset. The generated sewing patterns are located at examples). python3.9 .\data_generator\pattern_gen.py -c meta_infos/configs/dataset_config.yaml -o examples

  2. Generate posed smpl files with different pose sequences. path_to_maya2020\bin\mayapy.exe .\data_generator\fbx_anime.py -o examples\\posed_fbxs

  3. Simulate the garmetns with different poses (Please update the config file located in meta_infos\\configs\\data_sim_configs.json to ensure all the prepared data resource are correctly located). path_to_maya2020\bin\mayapy.exe .\data_generator\data_sim.py

Acknowledgement