I have been using openvino successfully on my machines for person tracking (
Question: Does Openvino support the Intel UHD Graphics 630 (which seems to be the standard "low end" GPU)?
That being said now when I see the pictures live, I see the "boxes" around the people when there are no people in the view. That is, I see people without boxes and then I see boxes floating around with no people (to mark the people - as if the system now sees ghosts).
Anyone encounter this before?
Thanks for reaching out.
OpenVINO does support the Intel UHD Graphics 630. The performance drop on the CPU is expected as the CPU is acting as a general-purpose computing device that handles multiple tasks at once. Meanwhile, a machine learning model's performance can be affected by any reason such as algorithms, input training data, etc.
My suggestion, try to use the FP16 precision model on GPU and see if you might be able to get higher FPS and lower latency results. Intel-integrated GPUs have native support for FP16 computation and therefore support FP16 Deep-Learning models quite well.
Besides, refer to Benchmark App for code examples of performance measurement.
This thread will no longer be monitored since we have provided information. If you need any additional information from Intel, please submit a new question.