Forums > MaxMSP

Kinect blobs detection -> GL objects problem.

April 4, 2012 | 6:14 am

Hello.

I am making a patch that generates GL shapes over blob’s coordinate from Kinect.
I did connect each blobs to GL shapes, but when multiple blobs (especially more than 2 blobs) moves from the inside to outside and side by side of Kinect screen, it remains a residual object at the end of both side of the jit.window screen.

I thought there should be a ‘clear’ message or something else that makes clean out the gl object which doesn’t have blob’s coordinate..

I have spent whole day to fix this issue, still no luck. I originally used poly~ object for multiple blob connection. this problem gave me a pain.
Can anyone finds a solution for this problem?

Please connect a Kinect and set it detects two or more solid objects’ blobs. Then swivel the Kinect horizontally.
You will see black gl bar objects follows the blobs, but when these blobs move inside and outside of screen, there’s a residual object at the center and both side of the jit.window screen.

please help.

– Pasted Max Patch, click to expand. –

April 4, 2012 | 5:20 pm

have you try to implement the "gate no blobs" from kvl patch to filter the "none blobs detection" ?

– Pasted Max Patch, click to expand. –

April 4, 2012 | 7:08 pm

yes, I have considered ‘no blob’ from KVL. but it appears cv.jit.blobs.centroids generates a list and individually connected to a gridshape. As the kinect doesn’t give bang when no blob detected, it seems like each blob’s list won’t be reset in GL when the number of blobs or order assignments are changed. Isn’t this problem belongs to GL processing, since the blob’s list gives 0,0,0 (no gl object) when there’s no blob for each assignment. Is it really need to use multiple blob tracking for generating GL shapes from blob lists?

I don’t know where this problem caused from.


Viewing 3 posts - 1 through 3 (of 3 total)