Amper

    r = range(ker_size)
    R = map(lambda i: map(lambda j: (i,j), r), r)
    for x, y in map(lambda x: map(lambda y: (x,y), im_size[1]), range(im_size[0])):
        #loop over channels
        for c in range(chs):
            acc = 0
            #loop over kernel
            a, b = y - ctr, x - ctr
            for i,j in R:
                n, m = a + j, b + i
                if 0 <= m < im_size[0] and 0 <= n < im_size[1]:
                    acc += data[m][n][c] * kernel[i][j]
            out_data[x][y][c] = acc