Develop

  1. Motivation
  2. Structure
  3. Start
    1. 1. Triangulate
    2. 2. Fit SMPL
    3. 3. SMPLify
    4. 4. Multiple Person

Motivation

  1. For beginners in the Mocap field, learn PyTorch programming by replicating the basic Mocap algorithms, and the basic principles of motion capture.
  2. For MoCap researchers, you can learn how to standardize the addition of modules and implement new algorithms in EasyMoCap through this document. Focus directly on the development of the algorithm, without needing to repeatedly implement input/output operations and visualization.
  3. For EasyMoCap users, you can preview some codes not included in the official version in this document.

Structure

Our code consists of three main parts:

Part  
dataset This part provides data for a specific time point, such as a single image or multiple images from different perspectives at that time point.
at_step This part processes the data at each time point. For example, it estimates keypoints for a single frame or triangulates keypoints for the current frame.
at_final This part performs post-processing on the entire dataset after iterating through all the data. This can include offline smoothing or fitting the SMPL model.

Start

We learn through a series of examples, each of which includes the corresponding data. You can download them and run the corresponding code.

1. Triangulate

Start learning from the triangulation of single-person 2D keypoints from multiple perspectives.

Body Hand

2. Fit SMPL

Fit a SMPL parametric model to 3D keypoints.

Model Output
SMPL
MANO

3. SMPLify

Fit SMPL model to monocular 2D keypoint input.

Model Output
SMPL
MANO
Fixed MANO

4. Multiple Person

Fit SMPL model to multiple person from multi-view inputs.

Model Output
Ballet
Soccer

Table of contents