Another weekly recap on the OpenCV Spatial AI Contest 2021 (Satya Mallick Phil Nelson) #OAK2021 from #KAUDA team (w/ Giovanni Lerda), this time focus back to #depthai #unity #plugin (Brandon Gilles) So this is a followup on previous post using #facedetection to fly an airplane. As a natural step forward, now it's time for 2-stage pipeline. The first step is the face detector and using the results for the second stage in this case head pose estimation (yaw, roll, pitch) ✅ If you're interested on use oak devices inside unity, take a look on the repo here 👉 https://lnkd.in/eqz5Af9 ✅ if you struggle to find examples like this one on C++, check the repo here 👉 https://lnkd.in/eQN6Awd #madewithunity #artificialintelligence #machinevision #diy #edgeai #edgecomputing #machinelearning #reinforcementlearning #deeplearning #digitaltwin #robotics
So cool 🙂
Nice!
Repo just updated with head pose example. All the info in the readme. https://github.com/gespona/depthai-unity-plugin
Just to clarify the head movement is not updated at same rate like the image preview you see ... it’s on propuse to avoid some small flick but could be updated at same frame rate like the image preview ... you could try it soon in the repo
Awesome! On the update rate in Unity - it looks a bit slower than the update rate of the head pose metadata results. Is that intentional to prevent jumpiness perhaps? Or is there some other limitation/reason?
It looks great!
Check out Rigging.js! Very similar but with TensorFlow.js in browser -very easy to use without any deps https://www.youtube.com/watch?v=OCoEztcBNms
So cool, my nephews will lose their minds when I send them this
Great job!!! congratulations
Data Engineer | Analytics Engineer | Generative AI Engineer
3yCool job!!