View source on GitHub |
Rotate image(s) counter-clockwise by 90 degrees.
tf.image.rot90( image, k=1, name=None )
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. |
© 2020 The TensorFlow Authors. All rights reserved.
Licensed under the Creative Commons Attribution License 3.0.
Code samples licensed under the Apache 2.0 License.
https://www.tensorflow.org/versions/r1.15/api_docs/python/tf/image/rot90