Django, Python OpenCV and Pycuda

Accelerated Computing GPU-Accelerated Libraries
#1

Hi! I’m working on a project that basically lets the user upload an image and do some opencv operations on them. However, as part of my requirements, I need to utilize the GPU in processing the images. My whole implementation is on Python (and Django) and the CUDA Accelerated functions in opencv are only available in C++. I am exploring Pycuda but I don’t know how can I call this gpu accelerated functions (e.g. cv::gpu::threshold(src, dst, 128.0, 255.0, CV_THRESH_BINARY) ) inside a kernel.

Any help would be appreciated. Thanks!

#2

I know the post is old but at the time it’s not possible i’m suffering alot to move my code from numpy, pillow, an opencv on python to the TX2 several libraries are not available, is a pain the lack of support.

#3

Had to reduce the scope of my study due to the lack of support. Hoping they’d provide better compatibility on python in future releases. Good luck on your project man.