Leave data_format as None in keras.ops.images and add data_format support to pad_images and crop_images

[james77777778]

As titled.
This should follow the convention of keras.layers:
Leave data_format as None, and if not specified, set it to the value of keras.config.image_data_format.

Additionally, the logic of pad_images and crop_images have been updated to support data_format. With this PR, we can also process unknown shapes in pad_images (similar to ops.pad).
However, it is impossible to support fully unknown shapes in crop_images.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 82.41758% with 16 lines in your changes are missing coverage. Please review.

Project coverage is 73.10%. Comparing base (a243d91) to head (aabc537).

Files	Patch %	Lines
keras/src/ops/image.py	81.69%	7 Missing and 6 partials ⚠️
keras/src/backend/torch/image.py	40.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #19774      +/-   ##
==========================================
- Coverage   78.56%   73.10%   -5.46%     
==========================================
  Files         498      498              
  Lines       45715    45742      +27     
  Branches     8440     8440              
==========================================
- Hits        35916    33440    -2476     
- Misses       8076    10654    +2578     
+ Partials     1723     1648      -75     

Flag	Coverage Δ
keras	73.02% <82.41%> (-5.39%)	⬇️
keras-jax	62.13% <75.82%> (+0.01%)	⬆️
keras-numpy	56.40% <78.02%> (+0.02%)	⬆️
keras-tensorflow	63.44% <73.62%> (+0.01%)	⬆️
keras-torch	?

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.

[fchollet]

LGTM, good call