Those who follow the footsteps of Microsoft Research, we are not surprised to find that they are working on improving the system and hand gesture recognition through cameras. The last thing that has come out of the oven is called Handpose.
Handpose is software associated with the system of recognition of Kinect, and as you can see in the videos that we share, it is very fast and recognizes in detail each element that makes up our hand.
Microsoft continues to seek new ways of complementary systems that appear interaction, and we can go beyond “screens”: Handpose also can get along with helmets of virtual reality. Initially its developers show us how it works in ordinary monitors, or superimposing the result on the video recording. Always with a Kinect unit recording the play.
The definition in fingers and hands is probably the most important to evolve in this kind of recognition element will allow us to catch elements in the distance, interact with more complex interfaces. Can you imagine a virtual paint box with your own hands? Defuse a bomb remotely? Recognize sign language? Handpose is on the path:
To begin using Handpose, the software creates a 3D model of our hands, which takes little more than one second to make. Once created, we can see on screen the instantaneous movement – 30 frames per second – each of our fingers and the rest of the hand.
They tell us that the system takes 30 points of reference that cover all of our hand and the real secret of all the invention are in the software and the ability to process, which multiplies by ten the usual recognition on Kinect. It will be more difficult to bring to the mobile world, but also one of its future applications.