Qualisys AIM Model improvement

[Martin Molloy]

Re: Qualisys AIM Model improvement

Hi Yakushevsky

The general advice is to include as great a range of movement as possible in a trial including movement similar to that you wish to capture. I get good results from recording 30 seconds of general movement at the same frame rate as the trials I wished to capture. I have yet to test if including motion throughout the capture volume improves the AIM prediction, perhaps someone else may comment on this.

Additionally, you may wish to consider the following:
Include data from a number of subjects if relevant.
Don't add data to the AIM model from poor quality captures (low percentage marker visibility, even if it's a low number of markers).
Make sure QTM is up to date as Qualisys is continually improving AIM.

Hope this is helpful.

[Serge Yakushevsky]

Hello,
In ourlaboratory we work with Qualisys (QTM system).
My question:how can I improve AIM model to reduce manual corrections of unlabeled markers?At the beginning of the measurements, I define AIM model on dynamic motion fileand then just adds new measurement to existing AIM model, exists other way ortechnic for resolving this problem?

Best Regards,
Yakushevsky Serge,
Ben-Gurion University,
Dep. of Industrial and Management Engineering