Alpha_I_Pixel concerns

[joaco18]

Hi! Thanks for your implementation of this paper it has been very helpful and I've learnt some cool efficiency tricks from it.

I have a doubt about the vectorized function Alpha_I_Pixel:

@jit(nopython=True)
def Alpha_I_Pixel(item,minVal, maxVal):
    return (item * minVal) / (maxVal - minVal)

I don't see in the paper where this operation is applied, in equation 15 it applies an operation equal to T_or_I_of_Neutrsophic function not Alpha_I_Pixel. Am I missing something?

Also (this is not an issue but a curious doubt), I've notice you do the mean filtering in a separate fashion. I can see that the mean filter is a separable one, so the two step convolution should be the same as doing it in only one step with the square filter. I was intrigued and checked the computation time, in deed doing it separately takes almost half the time, but I checked the results and they are not equal. I simply did the difference between the results from doing it in two steps or in one with the square filter and surprisingly I found that not all the results were the same, yours seems to return higher values. Maybe you've checked this and have some thoughts to share!

Thank you in advance!
Joaquín