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/r1.15/api_docs/python/tf/keras/preprocessing/image/random_rotation