wavenetr: Wavelet-Band Transformer Segmentation Networks

wavenetr provides TensorFlow/Keras WaveNETR models for 2D and 3D segmentation using WaveletViT bottleneck blocks.

Capabilities

• WaveNETR2D and WaveNETR3D segmentation models.
• WaveNETR2DSquareCore and WaveNETR3DSquareCore for fixed square/cubic inputs.
• CVWaveNETR2D and CVWaveNETR3D variational variants.
• Channel-last TensorFlow/Keras inputs.

Limitations

• Square inputs are required for WaveNETR2DSquareCore.
• Cubic inputs are required for WaveNETR3DSquareCore.
• Spatial sizes must be compatible with the configured wavelet levels and token stride.

Dependencies

• tensorflow
• waveletvit

Installation

pip install wavenetr

Minimal Example

import tensorflow as tf
from wavenetr import WaveNETR3D

model = WaveNETR3D(
    input_shape=(32, 32, 32, 1),
    n_classes=4,
)

x = tf.random.normal([1, 32, 32, 32, 1])
y = model(x)

print(y.shape)

Citation

This software is released for broad research, educational, and engineering use. If this package helps your work, please cite the following paper:

@misc{tarafdar2026interpretablefrugallearningsystems,
      title={Interpretable and Frugal Learning Systems Employing Multiresolution Pyramids and Volterra Kernels},
      author={Kishore Kumar Tarafdar},
      year={2026},
      eprint={2606.15011},
      archivePrefix={arXiv},
      primaryClass={eess.SP},
      url={https://arxiv.org/abs/2606.15011},
}

License

Apache License 2.0. See LICENSE.

Copyright 2026 Kishore Kumar Tarafdar.