tf.keras.preprocessing.image.random_rotation
Performs a random rotation of a Numpy image tensor.
tf.keras.preprocessing.image.random_rotation(
x, rg, row_axis=1, col_axis=2, channel_axis=0, fill_mode='nearest',
cval=0.0, interpolation_order=1
)
| Arguments | |
|---|---|
x |
Input tensor. Must be 3D. |
rg |
Rotation range, in degrees. |
row_axis |
Index of axis for rows in the input tensor. |
col_axis |
Index of axis for columns in the input tensor. |
channel_axis |
Index of axis for channels in the input tensor. |
fill_mode |
Points outside the boundaries of the input are filled according to the given mode (one of {'constant', 'nearest', 'reflect', 'wrap'}). |
cval |
Value used for points outside the boundaries of the input if mode='constant'. |
interpolation_order |
int, order of spline interpolation. see ndimage.interpolation.affine_transform |
| Returns | |
|---|---|
| Rotated Numpy image tensor. |
© 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/r2.4/api_docs/python/tf/keras/preprocessing/image/random_rotation