Hello everyone,
I’m working on a project built on Python 3.4 that uses OpenCV and OpenCV’s ArUco module for fiducial marker detection. We want to CUDA-accelerate our code, and I noticed that OpenCV 3 (which we are using) already has CUDA implementations of functions that the algorithm requires, such as threshold: http://docs.opencv.org/trunk/d8/d34/group__cudaarithm__elem.html#ga40f1c94ae9a9456df3cad48e3cb008e1.
If I am building OpenCV with the flag “WITH_CUDA=ON”, does this mean I am already accessing the CUDA-accelerated version of threshold when I call threshold?
cv2.threshold()
Or is it not possible to call the CUDA module functions from Python?
Thanks!