We’re already far in to February and it’s about time we reveal which two works we picked as winners for January. As we expected, the competition started of slow and we had to wait until the first week of February to have two proper entries. These came from Italian artist and teacher Michele Cremaschi and a group of researchers from the Humanoid Robotics Group at TU Bergakademie Freiberg.
Workflow and tutorial: Augmented Pinocchio
Michele explains how Augmented Pinocchio was created: “I produced a realtime hologram show where all the actors are hologram reproductions of my live acting; and in which all the props are holographic visuals generated in realtime relative to my body position. To generate the silhouette, I first used NI mate’s syphon one feed. In a second step, in order to have better resolution, I wrote a software that calibrates the Kinect with an external camera. This gives me a hires silhouette based on the depth mask syphon stream coming from NI mate.”
“Another point you might find interesting is how I networked the data coming from nimate; OSC runs easily on a network, but I also shared syphon streams thanks to Airparrot + airserver on another mac. This allowed NI mate streams to be used on a secondary computer, splitting the required CPU power (NI mate is nice, but quite CPU hungry) over two Macs.” Watch Michele’s tutorial and go to his blog to download the patch he created for extracting the skeleton data from NI mate in Vuo:
Workflow: Imitation learning for NAO robot using NI mate
Who doesn’t love robots – as hardware geeks we definitely do. In this super cool entry from researchers at the Humanoid Robotics Group at TU Bergakademie Freiberg, one a NAO robot is taught to imitate movements captured from real people. Oliver Mothes from the research group explains:
“We tried to use NI mate for human robot imitation, especially for motion tracking. The NI mate OSC interface help us to send the joint positions to our Blender-Robot-Skeletal. With a little bit magic, we transform these positions to robot joint angles, optimise them for stabilising the motion and send the new angles to the real robot.”
Thanks for entering and congratulations to the people behind these entries. We keep being amazed and super happy in seeing NI mate being used in such varied and creative ways. For the rest, there are still almost two weeks to post your entries and have a chance to win licenses of NI mate – as well as a full Oculus Rift devkit! Have fun!