Noise Estimate (MAD)

Problem

Estimate the noise standard deviation from horizontal neighbor differences.

Take the median of abs(I[i][j+1] - I[i][j]) over all adjacent pairs and divide it by 0.6745, rounding to 4 decimal places. An image with no horizontal pairs returns 0.0.

Test cases

case 1

in: [[[0,10,0],[10,0,10]]]

expected: 14.8258

case 2

in: [[[0,2,4]]]

expected: 2.9652

case 3

in: [[[5]]]

expected: 0

case 4

in: [[[7,7],[7,7]]]

expected: 0

Starter codePython

def noise_mad(image):
    # Your code here
    pass

4 test cases · run your code against them in the editor