addons

addons copied to clipboard

Currently, dense_image_warp uses only bilinear interpolation. It would be beneficial to use other interpolation schemes for better quality or different tradeoffs.

Other interpolation schemes are already available in other operations such as resize (Lanczos interpolation 3x3 or 5x5). tf.image.resize(in, new_size, method = tf.image.ResizeMethod.LANCZOS5)

Relevant information

• Are you willing to contribute it (yes/no): yes, see next point
• I created a repo containing and explaining a working implementation of the catmull-rom (quite similar to Lanczos-2) interpolation as well as a comparison and a test https://github.com/Luvideria/tensorflow-dense-warp-catmull-rom
• Are you willing to maintain it going forward? (yes/no): I don't know

Which API type would this fall under (layer, metric, optimizer, etc.) image

Who will benefit with this feature? Anyone using dense_image_warp and looking for more quality rather than speed.

To avoid duplications as we had also a request related to elastic image warping in Keras-cv can you try to open a ticket there?

You mean replying to the existing issue or raise a new issue?

As you like

TensorFlow Addons is transitioning to a minimal maintenance and release mode. New features will not be added to this repository. For more information, please see our public messaging on this decision: TensorFlow Addons Wind Down

Please consider sending feature requests / contributions to other repositories in the TF community with a similar charters to TFA: Keras Keras-CV Keras-NLP