为什么“imgaug Flipud”不适用于多边形？

Question

import imgaug as ia
import imgaug.augmenters as iaa
import cv2
import numpy as np
import matplotlib.pyplot as plt

# image is numpy array of size (327, 327, 3).
images = np.expand_dims(image, axis=0)

# define polygons
polygons = [
            [   ia.Polygon([(169, 173),
                            (140, 196),
                            (134, 225),
                            (142, 239),
                            (156, 250),
                            (174, 245),
                            (187, 230),
                            (187, 212),
                            (174, 197),
                            (165, 194),
                            (182, 178)]),

                 ia.Polygon([(149, 96), 
                             (164, 72), 
                             (183, 99),
                             (200, 93),
                             (171, 51),
                             (135, 92)])
            ]
        ] 


# define augmentor
seq = iaa.OneOf([
        #iaa.Affine(rotate=(-20, 20)),
        #iaa.Fliplr(1),
        iaa.Flipud(1)
])

# apply augmentation
images_aug, polygons_aug = seq(images=images, polygons=polygons)

# manipulate polygons obtained from augmentation
p1 = np.array(list(zip(polygons_aug[0][0].xx_int.tolist(), polygons[0][0].yy_int.tolist())))
p2 = np.array(list(zip(polygons_aug[0][1].xx_int.tolist(), polygons[0][1].yy_int.tolist())))

# combine polygons into list and draw onto the augmented image
p = [p1, p2]
cv2.drawContours(images_aug[0], p, -1, 255, 3)
plt.imshow(images_aug[0])

这里，augmentor给出了和原始坐标相同的多边形坐标。它适用于Fliplr和Affine，但不适用于Flipud。我也尝试过Flipud与边界框太，它是工作良好