I´m working with the Intel Perceptual Computing SDK and I my question is about the hand pose and gesture recognition (new ones, not the predefined ones). I know there are some related topics in this forum and I have read the manual, but I still haven´t understood clearly how to do it properly. Do I have to use the GeoNodes to track the fingers and use those data or is there a better way?
For example, if I would like my application to recognize a pose that is my hand with three finger extended, do I have to work with the data of the GeoNodes and compare distances or can I register new poses for using them as the predefined ones (lBIG_5, THUMB_UP...)?
Is there any tutorial about this? or do you know the proper steps I should follow for doing that?
PD: I´m using C++
Thanks for your help!