Project 3: Skeleton Generation from an Optical Motion Capture System

Project Team:

Jacob M. Kretz

Chunmei Lu

This project was broken into two parts: 

   First, we needed to generate a file parser to get the data from the text files.  This allowed us to then build a marker viewer.  This viewer has limited functionality, but did allow to see if these first steps were being done correctly.

   Second, we then wrote a complete BVH file generator.  This computes the skeleton hierarchy, the joint position between each set of bones, the orientation of the bones, and the lengths of the bones. The code for this and a further description of this is located on Chunmei web site.

Marker Viewer

The marker viewer program does two things.  It first parses the text file that contains the optical data.  There are certain assumptions made about this parser that are stated in the header files.

The second thing this program does is displays an user interface that allows the user to look at the marker data.  The markers are represented as cubes.  The user interfaces allows for the panning and rotating of the camera, the zooming of the camera, and the running of the complete motion.  The running of the complete motion must be done by dragging the slider as the run button does not work.  With this viewer the motions captured are definitely visible and recognizable.

Note:  The cubes that are not part of the motion are drawn because these particular values could not be interpolated.

CODE

File Parser Code

MotionCap.H

MotionCapInputParser.CPP

Drawing Class

DrawMarker.H

DrawMarker.CPP

User Interface Class

p33.H

p33.CPP

Main

MotionCapParse

Executable

MotionCapProject.exe

Presentation

CS838 Presentation