Numba, PyCUDA for GPU computing

I am new to this topic completely. My program is about finding defective pixels in screen. I have my program written in Python using many libraries of OpenCV. It runs fine on CPU but takes time and would like to get this running on GPU, I have GeForce GT 1080Ti installed .

I tried to work on PyCUDA but its difficult to write kernal for every opencv library being used.

Any suggestions how can i proceed further or start over?

OpenCV has a GPU accelerated version.

Any reference source please?
does it work with Python?

https://docs.opencv.org/2.4/modules/gpu/doc/introduction.html

https://stackoverflow.com/questions/42125084/accessing-opencv-cuda-functions-from-python-no-pycuda/52436378#answer-52436378

Thank you
I will get back in case of any futher queries

I have tried this OpenCV GPU accelerated version
http://www.decom.cefetmg.br/docentes/flavio_cardeal/Teaching/cvision/opencv3.pdf

but the time taken is still the same and i do not think this is utilising GPU