2

I'm trying to make a mobile app on image recognition(Computer Vision Application) . Does anyone know whether modern day smartphones have enough processing power/memory to recognize, say about 1 million classes from their real-time camera feed (30 fps) ?

(On-device inference or YOLO neural net concepts needed.)

What is the maximum number of classes it can reliably classify on a mobile device?

Some insight into figuring out computational loads/times in general would be helpful.

Brian Spiering
  • 20,142
  • 2
  • 25
  • 102
R. Rudinthkin
  • 21
  • 1
  • I can answer your very first question, a smartphone does NOT have the processing power to recognize 1M classes, at least not with any meaningful latency. For questions 2 & 3 you will have to do some testing yourself, I doubt many people have published any research/numbers on this. – Sauraus Jan 05 '19 at 17:18
  • Why would you want to discriminate a million classes? – S van Balen Feb 26 '20 at 15:13

0 Answers0