We have tested some very simple sequence using only one webcam. The result can be seen in http://vis.uky.edu/~jameszhao/mocap_ipisoft

I know it is still very rough as I have not fully exploited all the feature of your system and it only utilized one camera, but I do need some advices for conducting the future experiment.

The main problem in the sequences is the part when I cross my arms and tried to do something like a golf swing (sorry I don’t play golf so it looks very unprofessional~~), most of my questions are focused on this issue,
1. Do you think multiple camera will solve the “cross arm” problem? It seems to me that all the information is already in the frontal view and the view from other cameras may not help much for this specific problem
2. I guess one of the reason for the tracking failure is that the color of my sleeves are the same of the body clothes, so the system cannot really detect the arms. Using short sleeve shirt may solve this problem, but I fail to find any option in the software to allow people wearing short sleeve shirts. Am I right about this?
3. Video 2 has significant better tracking than video 1. The major difference between the two videos is that I slow down my movement. Alternatively, I can expedite the capture speed, but it is only possible by using industrial cameras. However, I fail to use the capture software for our 1394 cameras. Any recommendation on high speed capture?

sera q funciona mesmo eim??

This golf swing motion is too complex for single-camera system. You need to use multi-camera system for complex motions like golf.