Can anyone explain to me what are these for?

I'd be interested to know as well.

_________________

These are visualizations of various constraints used during tracking.

Collision Objects - used to ensure arms do not penetrate torso etc.

Mass Objects - used to calculate static/dynamic balance of the character.

Pose Mismatch - visualization of pose constraints imposed by video. Green color means good match. Red means bad match. Technically, markerless tracking means non-linear optimization with the goal of getting more green and less red.

Okay this sounds interesting because I always get this. So how is it used here? Do you mean before pressing track forward, I need collision objects selected too to prevent DMC from penetrating the body?

By the way, I heard about the recent terror attack in the train stations there in your country. Looks terrible, I hope you guys are safe there and condolences to those who were victimized by this tragedy.

You don't have to select anything for collision detection to work. It is always processed automatically.

You should not be seeing arms penetrating torso. Maybe you've seen hands penetrating torso? Hands don't have corresponding collision shapes as of yet. We recently moved collision detection from CPU to GPU. GPU collision detection is dramatically faster. This should allow us to add more collision shapes.

And thanks for your sympathy.

You are right. Looks like the arms doesn't penetrate the torso but the hand does.

http://yfrog.com/0epen1rjx

I know you guys can still improve DMC. Well, I should leave you guys to what you are doing. Thanks for the reply.