Deep Learning Computer Vision models depend on massive datasets for their advance computations. There are multiple human motion capture datasets available which serves the purpose of doing the base computations but lack to be the ideal dataset in two major ways. Either, the datasets are very small and are constrained to only particular motion or the output generated by these datasets is not natural, that is, it’s not true to our humane movements. These problems are solved by AMASS (Archive of Motion Capture as Surface Shapes) which generates highly versatile SMPL (Multi Person Linear Model) model outputs which includes standard skeletal representation and a full surface body mesh. The MoSh algorithm which is used to generate the output is modified to incorporate the soft tissue dynamics and renamed to MoSh++. The dataset is then made richer by adding data from datasets like DMPL (Dynamic SMPL) and MANO (Hand Model with Articulated and Non-rigid Deformations). AMASS is able to generate output which is compatible with recent graphics engines, game engines and animations requirements.
1. REQUIREMENTS:
Python 3.7 Human Body Prior Pyrender for visualizations
Install from this repository for the latest developments:
pip install git+https://github.com/nghorbani/configer
pip install git+https://github.com/nghorbani/human_body_prior
pip install git+https://github.com/nghorbani/amass>
2. Get the data: Download the body data .npz files from https://amass.is.tue.mpg.de/datasets or one can use any other body data in. npz format.
3. For the next steps, it is important to follow the exact steps in the Jupyter notebook.
4. For Visualization from Section 3.1, follow this walkthrough: https://scientist-sanyukta.medium.com/human-motion-detection-using-amass-a510f57dd7ff
REFERENCES:
AMASS. (n.d.). AMASS: downloads. Retrieved from AMASS:https://amass.is.tue.mpg.de/users/sign_in?locale=en
D. Anguelov, P. S. (2005). SCAPE: Shape Completion and Animation of. ACM Transactions on Graphics,.
Georgios Pavlakos, V. C. (n.d.). Expressive Body Capture: 3D Hands, Face, and Body from a Single Image. MPI for Intelligent Systems.
Javier Romero*, D. T. (n.d.). MANO: Embodied Hands. Retrieved from MANO: https://mano.is.tue.mpg.de/
Javier Romero, D. T. (2017). Embodied Hands: Modeling and Capturing Hands and Bodies Together. ACM Transactions on Graphics.
LLC, 3. (n.d.). 4D Scan. Retrieved from 4D Scan: http://www.3dmd.com/
Matthew Loper, N. M. (33(6):220:1–220:13). MoSh: Motion and Shape Capture from Sparse Markers. ACM Trans. Graph, Nov. 2014.
Matthew Loper, N. M.-M. ( 2015). SMPL: A Skinned. ACM Trans. Graphics.
Naureen Mahmood, N. G. (n.d.). AMASS: Archive of Motion Capture as Surface Shapes. MPI for Informatics.
nghorbani. (n.d.). Github. Retrieved from human body prior:https://github.com/nghorbani/human_body_prior
Nvidia. (n.d.). CUDA Developer. Retrieved from CUDA Developer:https://developer.nvidia.com/cuda-zone
SMPL. (n.d.). SMPL eXpresive. Retrieved from SMPL eXpresive: https://smpl-x.is.tue.mpg.de/downloads