Benchmark existing algorithms in the current python implementation on most heavy computational costs.
I assume that the rgb2hsv / sobel filtering costs the most, but we need to get some numbers to see what costs the most to minimize the optimization work.
It would be good to benchmark on pc and on the embedded platform (RPI3 e.g.).
Possible useful tools for this task:
Benchmark existing algorithms in the current python implementation on most heavy computational costs. I assume that the rgb2hsv / sobel filtering costs the most, but we need to get some numbers to see what costs the most to minimize the optimization work. It would be good to benchmark on pc and on the embedded platform (RPI3 e.g.). Possible useful tools for this task: