I thought I had it trained properly, but thanks for the link. There are so many competing sources of documentation for AR, it's a very noisy process trying to figure out which is the right one. (The one you sent seems to be the right one, at last)
I decided to do this project in C++, so I ported the fluxus ARToolKitPlus addon to Cinder: http://pixelist.info/artoolkitplus-in-cinder/ (I have a few questions on that, if anyone wants to take a peek) I would have like to have done it in fluxus, but I'd rather use the C++ Animata code for now due to time constraints. Still, I will be looking more at fluxus and AR, it was more fun than I thought! Best Evan On 21 Dec 2010, at 21:42, gabor papp wrote: >> Still, I'm using template mode (as per your example), and all I changed >> was the hiro.patt file (I replaced it with one of my own). It seems to > did you train the system properly to detect your marker? > if you don't, check out this part from the artoolkit documentation. > http://www.hitl.washington.edu/artoolkit/documentation/devmulti.htm > > does the hiro marker return an id? > >> work, because it detects it, but the pattern id is always #f. Should the >> id returned in pattern mode be the pattern id, and not a unique marker id? > if you load the trained pattern file from the .patt file, ar-load-pattern > returns an id. if the marker is recognised this id will be returned. > > best, > gabor
