Fixed missing axis annotation in tf map_coordinates

[Phil2852]

When calling the tensorflow version of keras.ops.image.map_coordinates with coordinates containing an index outside of the input image dimensions and constant padding, the output image is filled with fill_value due to a missing axis annotation in the tensorflow backend code.

Script to reproduce with current master branch:

import keras
import tensorflow as tf
import matplotlib.pyplot as plt


def main():
    img = tf.random.normal((128, 128))
    # identity map for comparison
    grid = tf.stack(
        tf.meshgrid(
            tf.range(0, img.shape[0], dtype=tf.float64),
            tf.range(0, img.shape[1], dtype=tf.float64),
        ),
        axis=0,
    )
    # first row and column are -1 and therefore considered out-of-bounds
    grid_with_oob = tf.stack(
        tf.meshgrid(
            tf.range(-1, img.shape[0] - 1, dtype=tf.float64),
            tf.range(-1, img.shape[1] - 1, dtype=tf.float64),
        ),
        axis=0,
    )

    identity_mapped = keras.ops.image.map_coordinates(img, grid, 1)
    mapped_with_oob = keras.ops.image.map_coordinates(img, grid_with_oob, 1)

    fig, ax = plt.subplots(1, 3)

    ax[0].imshow(img.numpy())
    ax[1].imshow(identity_mapped.numpy())
    ax[2].imshow(mapped_with_oob.numpy())

    plt.show()


if __name__ == "__main__":
    main()

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 82.60%. Comparing base (dfaca0e) to head (1bf1eb3).
Report is 13 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #21304      +/-   ##
==========================================
+ Coverage   82.57%   82.60%   +0.02%     
==========================================
  Files         564      565       +1     
  Lines       54677    54772      +95     
  Branches     8500     8508       +8     
==========================================
+ Hits        45152    45243      +91     
- Misses       7435     7439       +4     
  Partials     2090     2090              

Flag	Coverage Δ
keras	82.41% <100.00%> (+0.02%)	⬆️
keras-jax	63.55% <0.00%> (-0.06%)	⬇️
keras-numpy	58.71% <0.00%> (-0.04%)	⬇️
keras-openvino	33.13% <0.00%> (+0.08%)	⬆️
keras-tensorflow	63.96% <100.00%> (-0.07%)	⬇️
keras-torch	63.62% <0.00%> (-0.06%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[fchollet]

Thanks for the PR! Please add a unit test for the fix.

[Phil2852]

Sure, I added a rather generic constant padding test that would have caught the issue.

[fchollet]

LGTM, thank you for the fix.