Blob tracking

Hi all, first i wanna tell u ppl that vvvv is really nice project, like the way energy works around it :)

I come from Eyesweb/opneCv/reaktor usage…I build up some installation/projects, if somebody is inerested i’ll post links.

Anyway i was thinking to reassemble big pieces of software using vvvv, so i’m considering what and how can i do, first of all os there any exapmles of blob traking ? which method is the best for it according to vvvv code ?

I looked in forum and read some thread but i didn’t found a sample to understand logic behind it, if somebosy as hints or whatever i’ll appreciate.

Thank u all in advance and sorry for too much words…is just the night

Ok, i’m happy…i saw “contour” and it’s using OpenCV :)

what the idea about an adaptive backgroundSubtraction in vvvv , how can be done ?

well…seems i’ll have very busy time :)

It’s not for multitouch, if u browse select products, u’ll see what i’m working on, anyway, these time i’m building a blob detection for an installation and i was thinking about use adaptive background as feature for changing-lights location.

BTW that’s a feature, i’m replying a kind of space where blob are acting like magnetic emitter, according to reaktor audio rendering and bypassing flex for video rendering

I mean, till these morning my tools were other, i’m trying to understand if and how change lots of stuff…goin to check ur hints, thank u: )

My Italian is very bad, but wow, nice webiste!!
That kilyka video is a human sound sequencer? COOL!! :)

Learning vvvv and build a project like that is not the easiest thing to start with. But, if you have Skype, you can get an Army off people who can help you out if you have some specifc questions.

edit: For Skype contacts.

Just wanted to say : nice stuffs…I like the way this sounds so like Reaktor :)

i’m very happy u ppl liked that, btw i’m gonna register to skype now before fainting on couch :)

Kilyka is a nice project…and actually uses human body cynematic structure as basis, then a lot of physic and geometrical stuff…anyway what’s nice is the way we can just interpretate a body with music and video (there is a video rendering too).

I think that translate all inside vvvv is a big work, but it’s a kind of porting, eavy but possible :)

Then other project KIS is very interesting too

But what i’m really burning byt since this morning is to develop installations with vvvv for several reasons (opencv dll for first)

It’s just I need to come into vvvv logic :)

West : -BG+key is not doing an adaptive bgSubtraction, but a masking using static bgSubtraction(with bg setted at beginning), thought was helpful to understand better vvvv.

For the blob tracking contour it’s eanough, i was just wondering if there was a module or a trick to adapt background…logically is like setting a new bg every ‘interval’, but i think it’s faster to check just changed pixels, i’m not a such developer or programmer, but i’ll understand how to do it :)

goin to skype stuff, thank u ppl, and i’m ready and happy for everykind of cooperation ;)

hello hierro, i would be interested in implementing an algorithm for adaptive background subtraction in vvvv. if you have links or papers about it, post them…

well, lately i hired a c++ developer to publish a DLL with some algo and once he was showing me AdptiveBgSub , but so far i think is a function or a group or function already implemented in OpenCV.

I took a look:

OpenCV Manual > see chapter 2


More theory

Workshop on specific purposes

A class i found in forum :D

The theory can be helpful, thus some application ar already using this kind of algo, i guess the best choice for general purpose should be based on shadows recognition, also to be used on projection, like this :

vvvv artist :)

(is the same method some german product like poolsystem uses)

Let me know what do u think about it, and thank u for interest