salvadorpsartek

I use a detection movement patch I made using cv.jit.

for each frame, I receive a list of between 20 and 40 couples (x,y).

what I need to do is to create average points with those points (that I receive from the movement detection).

I explain, if 3 points (2, 4, 5 or more) are close to each other (x2-x1

I create an average point (x1+x2+x3/3; y1+y2+y3/3).

I know it looks quite simple, but I can't reach what I want..

If anyone have an idea which could help me to solve it (or objects, externals), it would be much appreciated!

I use a detection movement patch I made using cv.jit.
for each frame, I receive a list of between 20 and 40 couples (x,y).
what I need to do is to create average points with those points (that I receive from the movement detection).
I explain, if 3 points (2, 4, 5 or more) are close to each other (x2-x1
I create an average point (x1+x2+x3/3; y1+y2+y3/3). 

I know it looks quite simple, but I can't reach what I want..
If anyone have an idea which could help me to solve it (or objects, externals), it would be much appreciated!
Many thanks,
J.


conditionned-average-points

not sure, but maybe something like this will get you started:

You might want an [abs] in between each of the [- 0.] and [

I've made an patch to detect changes in an determined position of an video window.

I used some "average engine" to let the patch knowing the suns position and changes in the light, to trigger only with heavy changes (like some one walking in front of the camera).

Copy the folder to your desktop and open demo.maxpat

To use with an list with two elements (x, y) you just have to duplicate the "blue backgound" patch.

I used some "average engine" to let the patch knowing the suns position and changes in the light, to trigger only with heavy changes (like some one walking in  front of the camera).

Conditionned Average points