I’m attempting to run some basic face detection on a Jetson TX2 using a USB camera and OpenCV. I used the provided OpenCV examples and found them to be extremely slow.
I used the profiling tool and see that the majority of the CPU time is spent in the Cascade filters, so I tried to change those for their CUDA versions. Unfortunately the included OpenCV module doesn’t seem to have the CUDA object detection or the CUDA resizing libraries.
Surely I am missing something here? How do we leverage the Jetson hardware in OpenCV?