Thanks Pedro :) I´m really not happy because I always end up running to many things that could be more easily integrated. When it comes do debugging, I really hate having to check out so many differente software. As I said before, it feels a little bit ridiculous to run Max/MSP and PD at the same time. I know, most things I could just write in Max/MSP, but then I wouldn't have flexibility i have in PD. That was a big issue during my master thesis.
I´ll check into those projects. Btw, I like OpenCV :) Best, Leandro On Mon, Dec 17, 2012 at 1:46 PM, Pedro Lopes <[email protected]> wrote: > They are all different things. ANN, HMM, etc... this is machine learning, > reasoning algorithms, and so forth. > > As far as I know Pd has no basic set of machine learning techs, such as > openCV has. > > I'm not a huge fan of replicating tools, so I would take existing external > tools and just interface with them, but since you point out that you are > not happy with them, my suggestion is useless. Anyway two pointers: > - http://sourceforge.net/projects/cvhmm/ (HMM) > - http://opencv.willowgarage.com/documentation/cpp/neural_networks.html(NN) > > Best, > pedro > p.s.: this message is not opencv sponsored. > > On Mon, Dec 17, 2012 at 11:50 AM, Leandro da Mota Damasceno < > [email protected]> wrote: > >> Thanks, Marco. I have been using OSC communications with other software, >> but I'm far from satisfied. I'll try ANN, thought. >> >> I am actually interested in working with Hidden Markov Models (and >> machine learning in general) and some filters for computer vision. Can I >> work with those using ANN? >> >> Best, >> >> Leandro >> >> >> On Sun, Dec 16, 2012 at 11:23 PM, Marco Donnarumma <[email protected]>wrote: >> >>> Hey, >>> >>> I've been using the ANN library. >>> I'm not sure it's still updated nor mantained these days, but it does >>> work well for my purpose. >>> >>> http://puredata.info/search?SearchableText=ann >>> >>> That is, I've been using it to make my instrument (the Xth Sense [1]) >>> detect a performer's muscle states throughout a piece by using 1 sensor >>> only; then, labelling different states, such as still, motion, fast motion, >>> slow motion. It's a basic implementation but useful to create a sensing >>> timeline that changes with the performer behaviours, rather than with fixed >>> time cues. >>> >>> Surely, there is much more to AI which cannot be presently done in Pd, >>> at the best of my knowledge, I might be wrong though. >>> Ben (Bogart) is our guy in this area. >>> >>> What are you using Ben? >>> >>> [1] http://res.marcodonnarumma.com/projects/xth-sense/ >>> >>> >>> >>> Hi all >>>> >>>> I have been wondering... Is there any AI implementation for PD? What >>>> have you been using for it? >>>> >>>> Best >>>> >>>> Leandro >>> >>> -- >>> Marco Donnarumma >>> New Media + Sonic Arts Practitioner, Performer, Teacher, Director. >>> Embodied Audio-Visual Interaction Research Team. >>> Department of Computing, Goldsmiths University of London >>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>> Portfolio: http://marcodonnarumma.com >>> Research: http://res.marcodonnarumma.com >>> Director: http://www.liveperformersmeeting.net >>> >>> >> >> _______________________________________________ >> [email protected] mailing list >> UNSUBSCRIBE and account-management -> >> http://lists.puredata.info/listinfo/pd-list >> >> > > > -- > Pedro Lopes (HCI Researcher / MSc) > contact: [email protected] > website: http://web.ist.utl.pt/pedro.lopes / > http://pedrolopesresearch.wordpress.com/ | > http://twitter.com/plopesresearch >
_______________________________________________ [email protected] mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
