Bilateral Filter

Problem

Apply a bilateral filter with a fixed 3x3 window.

For each pixel, weight each in-bounds neighbor by exp(-(di^2 + dj^2) / (2 * spatial_sigma^2)) * exp(-(diff^2) / (2 * range_sigma^2)), where diff is the intensity difference from the center pixel, then output the weighted mean rounded to 4 decimal places.

Hint

Multiply a spatial Gaussian weight by a range Gaussian weight for each neighbor.

Test cases

case 1

in: [[[7,7,7],[7,7,7],[7,7,7]],1,1]

expected: [[7,7,7],[7,7,7],[7,7,7]]

case 2

in: [[[0,0,0],[0,9,0],[0,0,0]],1,1]

expected: [[0,0,0],[0,9,0],[0,0,0]]

case 3

in: [[[0,0],[0,9]],1,1]

expected: [[0,0],[0,9]]

case 4

in: [[[0,0,0],[0,10,0],[0,0,0]],1,5]

expected: [[0.22,0.2708,0.22],[0.2708,6.5467,0.2708],[0.22,0.2708,0.22]]

Starter codePython

def bilateral_filter(image, spatial_sigma, range_sigma):
    # Your code here
    pass

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