peterzhousz / mocgui

Matlab Motion Capture GUI

Introduction

This page contains software and instructions for a GUI interface for viewing motion capture data.

Installation

1. Download the code via git clone https://github.com/zhfe99/mocgui.git or from this link;
2. Set mocgui/ as the current folder in Matlab;
3. Run addPath in Matlab to add sub-directories into the path of Matlab.
4. Run mocgui or demoMoc in Matlab.

Instructions

The package of mocgui.zip contains the following files and folders:

• ./data: This folder contains a subset of the CMU Motion Capture dataset.
• ./src: This folder contains the main implementation of the GUI interface.
• ./lib: This folder contains some necessary library functions.
• ./addPath.m: Adds the sub-directories into the path of Matlab.
• ./mocgui.m: The interface function.
• ./mocgui.fig: The Matlab fig file to save the window configuration.
• ./demoMoc.m: A demo file for loading and visualizing motion capture data.

FAQs

How to load more data?

• All the mocap data are located under data/cmu. In this package, I have put a folder data/cmu/S02 containing all the sequences performed by subject S02. You can download more data from the website of the CMU Motion Capture dataset and put the files in a similar structure as the folder data/cmu/S02.

How to show one frame of a mocap data?

• At the end of demoMoc.m, you can find an example of showing one frame of a mocap data.

How to get the coordinate of a mocap data?

• The function cmuMoc creates a structure variable named wsMoc, which contains all the mocap data for one sequence. In particular, wsMoc.QC is a 3 x nJ x nF matrix, containing the 3D coordinate of nJ joints of a nF-length motion capture data. You can check the code demoMoc.m for the example.

Copyright

This software is free for use in research projects. If you publish results obtained using this software, please use this citation.

@article{ZhouDH13,
    author    = {Feng Zhou and Fernando {De la Torre} and Jessica K. Hodgins},
    title     = {Hierarchical Aligned Cluster Analysis for Temporal Clustering of Human Motion},
    journal   = {IEEE Transactions Pattern Analysis and Machine Intelligence (PAMI)},
    year      = {2013},
    volume    = {35},
    number    = {3},
    pages     = {582-596},
}

If you have any question, please feel free to contact Feng Zhou ([email protected]).