Doubt when calculating mBIOU

[YZsZY]

Hello author, I think there are some problems in calculating the mBIOU.

def mask_to_boundary(mask, dilation_ratio=0.02):
    h, w = mask.shape
    img_diag = np.sqrt(h ** 2 + w ** 2)
    dilation = int(round(dilation_ratio * img_diag))
    if dilation < 1:
        dilation = 1
    # Pad image so mask truncated by the image border is also considered as boundary.
    new_mask = cv2.copyMakeBorder(mask, 1, 1, 1, 1, cv2.BORDER_CONSTANT, value=0)
    kernel = np.ones((3, 3), dtype=np.uint8)
    new_mask_erode = cv2.erode(new_mask, kernel, iterations=dilation)
    mask_erode = new_mask_erode[1: h + 1, 1: w + 1]
    # G_d intersects G in the paper.
    return mask - mask_erode

The dilation_ratio is set as 0.02, for a image with Height 700 X Width 900，the dilation iterations will be 24! The final mask will be an empty mask as shown below:

The binary you get at the end is actually the same as the input Mask (with erode 20x->empty mask), then the calculated mBIOU is actually the same as the mIOU.
Looking for your reply~