Honestly man, if we're not dealing with massive matrixes I don't think we really need to worry about time. Optimized code is good, but I've found myself spending literally dozens of hours trying to optimize code on the off chance that it's going to matter, just to realize that it, in fact, never mattered. If you have something nice and simple, and it works within the time constraint that we're working with, then keep it that way. We are using python after all.

This code looks elegant.
However, I checked it in Pycharm using timeit module and it is 3 times slower than just simple mapping through array using [row][col] indexes.

Way too many unnecessary actions for the code that can be written in one line and more effectively.