3 ms·
The "numpy" Sobel code is not that good, unfortunately - all the iteration is done in Python, so there is not much benefit from involving numpy. If one would us
by jononor 1y ago
The "numpy" Sobel code is not that good, unfortunately - all the iteration is done in Python, so there is not much benefit from involving numpy.
If one would use say scipy.convolve2d on a numpy.array, it would be much faster.