Rotate image(s) counter-clockwise by 90 degrees.
tf.image.rot90(
image, k=1, name=None
)
For example:
a=tf.constant([[[1],[2]],[[3],[4]]])
# rotating `a` counter clockwise by 90 degrees
a_rot=tf.image.rot90(a,k=1) #rotated `a`
print(a_rot) # [[[2],[4]],[[1],[3]]]
Args:
image: 4-D Tensor of shape [batch, height, width, channels] or 3-D Tensor
of shape [height, width, channels].
k: A scalar integer. The number of times the image is rotated by 90 degrees.
name: A name for this operation (optional).
Returns |
|---|
A rotated tensor of the same type and shape as image.
|
Raises |
|---|
ValueError
|
if the shape of image not supported.
|
TensorFlow 1 version
View source on GitHub