Joint Entropy of Image Pairs

Problem

Compute the joint Shannon entropy in bits of two same-shaped integer images.

Build the histogram of (a[i][j], b[i][j]) value pairs on the aligned pixels, then return -sum(p * log2(p)) rounded to 4 decimal places. An empty pair set returns 0.0.

Test cases

case 1

in: [[[0,0],[1,1]],[[0,0],[1,1]]]

expected: 1

case 2

in: [[[0,1],[0,1]],[[0,0],[1,1]]]

expected: 2

case 3

in: [[[0,1],[2,3]],[[0,1],[2,3]]]

expected: 2

case 4

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

expected: 0

Starter codePython

def joint_entropy(a, b):
    # Your code here
    pass

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