Deep Tracking Algorithms for Jetson tx1

Hi
I want to use a Tracking Algorithm with good accuracy and best performance.
I tried GOTURN Algorithm with opencv c++ on jetson tx1
but does not have good accuracy and I get 700ms FPS which is not good for real-time tracking.
How can I run GOTURN on GPU in C++ for the best performance?
What other deep-tracking algorithms can I try on Jetson tx1 with C++?

Hi,

Which framework do you use? Is it PyTorch?
If yes, you can try to deploy it with the GPU directly.

https://pytorch.org/tutorials/beginner/basics/quickstart_tutorial.html#creating-models

Thanks.

I use C++ and tensorrt for loading yolov5 model and also use opencv for using goturn

Hi,

Have you tried to boost the device’s performance?

$ sudo nvpmodel -m 0
$ sudo jetson_clocks

Based on the link below, GOTURN have GPU support already:

Thanks.

Hi
Thank you for your response
no I don’t have tried boost the device’s performance.
I Try the GOTURN GPU Support and let you know the result.
Thanks again .

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.