比较与tf.image.rot90的功能差异

tf.image.rot90

tf.image.rot90(
    image,
    k=1,
    name=None
)

更多内容详见tf.image.rot90。

mindspore.dataset.vision.Rotate

class mindspore.dataset.vision.Rotate(
    degrees,
    resample=Inter.NEAREST,
    expand=False,
    center=None,
    fill_value=0
)

更多内容详见mindspore.dataset.vision.Rotate。

使用方式

TensorFlow：逆时针旋转图像，每次旋转90度。

MindSpore：逆时针旋转图像，旋转角可为任意度数，并对旋转图像之外的区域进行像素填充。

代码示例

# The following implements Rotate with MindSpore.
import numpy as np
import mindspore.dataset as ds

image = np.array([[[0.1], [0.2]], [[0.3], [0.4]], [[0.5], [0.6]]])
result = ds.vision.Rotate(90)(image)
print(result)
# [[[0. ], [0. ]],
#  [[0.4], [0.6]],
#  [[0.3], [0.5]]]

# The following implements rot90 with TensorFlow.
import tensorflow as tf

image = tf.constant([[[0.1], [0.2]], [[0.3], [0.4]], [[0.5], [0.6]]])
result = tf.image.rot90(image, k=1)
print(result)
# [[[0.2], [0.4], [0.6]],
#  [[0.1], [0.3], [0.5]]]