From a novice can anyone say what the bottle necks for real time video inference?

Intel® Distribution of OpenVINO™ Toolkit

Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.

From a novice can anyone say what the bottle necks for real time video inference?

577 Views

OK, I have been playing with almost every sample that was included with a Pi 3B+ and

2 Movidius compute sticks. My basic question is, how do you get better performance?

According to the CPU monitor the PI cpu is barely taxed.

Its at 18 to 20%. So what changes can be made to get you the most bang for your inference buck.

Another board with USB3 like the ODroid XU4? Or maybe the Intel Up Board?

Maybe a different model? Maybe image size?

I am currently running at 640x480 but think 1/2HD 960x540 would deliver better edges.

Is it because I am using Python? Is there a guide to Movidius 30fps inference?

Also notice how this person is walking and not being detected. Is that because of the physical size?

https://ibb.co/gKdSf8

So far I have been able to track up to 30ft away with no real issue.

Link Copied

0 Replies

Community support is provided during standard business hours (Monday to Friday 7AM - 5PM PST). Other contact methods are available here.

Intel does not verify all solutions, including but not limited to any file transfers that may appear in this community. Accordingly, Intel disclaims all express and implied warranties, including without limitation, the implied warranties of merchantability, fitness for a particular purpose, and non-infringement, as well as any warranty arising from course of performance, course of dealing, or usage in trade.

For more complete information about compiler optimizations, see our Optimization Notice.