Rotate Arbitrary Angle

Problem

Rotate a 2D image by an arbitrary angle using nearest-neighbor sampling.

For each output pixel (i, j) let dy = i - (H - 1) / 2 and dx = j - (W - 1) / 2, then sample the input at round((H - 1) / 2 + dy * cos(a) + dx * sin(a)) and round((W - 1) / 2 - dy * sin(a) + dx * cos(a)), where a = radians(angle). Out-of-bounds samples become 0 and the output keeps the original shape.

Hint

Use the inverse rotation to map each output pixel back to a source pixel.

Test cases

case 1

in: [[[1,2,3],[4,5,6],[7,8,9]],0]

expected: [[1,2,3],[4,5,6],[7,8,9]]

case 2

in: [[[1,2,3],[4,5,6],[7,8,9]],180]

expected: [[9,8,7],[6,5,4],[3,2,1]]

case 3

in: [[[1,2,3],[4,5,6],[7,8,9]],90]

expected: [[3,6,9],[2,5,8],[1,4,7]]

case 4

in: [[[1,2,3],[4,5,6]],90]

expected: [[3,3,0],[1,1,0]]

Starter codePython

def rotate_nearest(image, angle):
    # Your code here
    pass

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