C3I SYNTHETIC HUMAN DATASET

Citation Author(s):
Shubhajit
Basak
Faisal
Khan
Hossein
Javidnia
Rachel
McDonnell
Michael
Schukat
Peter
Corcoran
Submitted by:
Shubhajit Basak
Last updated:
Mon, 07/08/2024 - 15:59
DOI:
10.21227/f6zx-bf29
Data Format:
Research Article Link:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

The C3I Synthetic Human Dataset provides 48 female and 84 male synthetic 3D humans in fbx format generated from iClone 7 Character creator “Realistic Human 100” toolkit with variations in ethnicity, gender, race, age, and clothing. For each of these, it further provides the full-body model with five different facial expressions – Neutral, Angry, Sad, Happy, and Scared. Along with the body models, it also open-sources a data generation pipeline written in python to bring those models into a 3D Computer Graphics tool called Blender. This framework, along with the virtual human models, can generate extensive synthetic facial datasets (e.g., Head Pose, Face Depths) with a high degree of control over facial and environmental variations such as pose, illumination, and background. Such large datasets can be used for improved, targeted training of deep neural networks.

Instructions: 

Following are the instruction to process the files in Blender -

  • Each folder under the expression folder containes the fbx file and the corresponding material files.
  • Open Blender (https://blender.org) and import the fbx file.
  • Add camera and lights to the scene and position them accordingly
  • Add composting layers in blender to get the ground truth files

We can also use the python api provided by Blender to generate the ground truth files directly running the scripts as provided here.


More details can be found in the following Github repository-

https://github.com/shubhajitbasak/blenderDataGeneration

If you find this work or dataset useful please cite the following work -

S. Basak, H. Javidnia, F. Khan, R. McDonnell and M. Schukat, "Methodology for Building Synthetic Datasets with Virtual Humans," 2020 31st Irish Signals and Systems Conference (ISSC), 2020, pp. 1-6, doi: 10.1109/ISSC49989.2020.9180188.

Funding Agency: 
Science Foundation Ireland Centre for Research Training in Digitally-Enhanced Reality (d-real) under Grant 18/CRT/6224

Documentation

AttachmentSize
File C3I_Synthetic_Human_Dataset.pdf252.26 KB