I’m the developer of a game that uses gesture recognition using the VR headset for HTC Vive, and I’m trying to improve the accuracy of gesture recognition.
(Game for context: http://store.steampowered.com//app/488760 . This is a game in which you cast spells by drawing characters in the air.)
I am currently using a $ 1 algorithm to recognize a two-dimensional gesture and using a spelling camera attached to the horizontal rotation of the player to smooth out the gesture that the player draws in space.
However, I am sure that there should be more effective approaches to the problem!
I should represent the gestures in 2D in the instructions, so ideally I would like to:
Find the optimal vector on which to smooth the gesture. Flatten it in 2D space. Use the best gesture recognition algorithm to recognize what kind of gesture it is.
It would be very helpful to approach 100% accuracy under any circumstances. Currently, for example, the game tends to get confused when players try to draw a circle in the midst of a battle, and it is assumed that they draw a Z shape instead.
All suggestions are welcome. Thank you in advance.
source share