I'm trying to figure out what's the best way to get a trained model that does facial detection to run on a camera that's connected to a computer and perform inference.

Since I don't need the device to train the model, is there a low cost solution that can still do at least 10fps on a 1280x720 image?

Is there any device that can run the model and have camera and computer all built in a small device?

If not, any suggestion on what's the best way to put one together and deploy it in production environment?



Since you said detection, you could do this on OpenCV: HAAR Cascade Detection for faces. It's swift, and should fit your use case. I've seen such scripts running on small Tello drones and gets job done relatively well.

Definitely there is hardware built around for this purpose. Look for Nvidia Jetson Nano and Jetson TK1/TX1. It's got sufficient juice to take even 30fps IMO.

  • $\begingroup$ Hi, thank you for your answer. However, the purpose of my post is to identify whether or not there's existing hardware that have good enough configuration which can do what I indicated above? $\endgroup$ – WorldWind May 4 '19 at 1:57
  • $\begingroup$ Also, Intel has their Neural Compute Stick 2 $\endgroup$ – Brale May 31 '19 at 7:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.