ptwt 0.0.3

Differentiable and gpu enabled fast wavelet transforms in PyTorch

Pytorch Wavelet Toolbox (ptwt)

Welcome to the PyTorch (adaptive) wavelet toolbox. This package implements:

• the fast wavelet transform (fwt) (wavedec)
• the inverse fwt (waverec)
• the 2d fwt wavedec2
• the inverse 2d fwt waverec2.
• single and two-dimensional wavelet packet forward transforms.
• adaptive wavelet support (experimental).
• sparse matrix fast wavelet transforms (experimental).

Installation

Install the toolbox via pip or clone this repository. In order to use pip, type:

$ pip install ptwt

You can remove it later by typing pip uninstall ptwt.

Example usage:

import torch
import numpy as np
import pywt
import ptwt  # from src import ptwt instead if you cloned the repo instead of using pip.

# generate an input of even length.
data = np.array([0, 1, 2, 3, 4, 5, 5, 4, 3, 2, 1, 0])
data_torch = torch.from_numpy(data.astype(np.float32))
wavelet = pywt.Wavelet('haar')

# compare the forward fwt coefficients
print(pywt.wavedec(data, wavelet, mode='zero', level=2))
print(ptwt.wavedec(data_torch, wavelet, mode='zero', level=2))

# invert the fwt.
print(ptwt.waverec(ptwt.wavedec(data_torch, wavelet, mode='zero', level=2), wavelet))

Unit Tests

The tests folder contains multiple tests to allow independent verification of this toolbox. After cloning the repository, and moving into the main directory, and installing tox with pip install tox run:

$ tox -e py

Adaptive Wavelets (experimental)

Code to train an adaptive wavelet layer in PyTorch is available in the examples folder. In addition to static wavelets from pywt,

• Adaptive product-filters
• and optimizable orthogonal-wavelets are supported.

Sparse-Matrix-multiplication Transform (experimental).

In addition to convolution-based fwt implementations matrix-based code is available. Continuing the example above try:

# forward
coeff, fwt_matrix = ptwt.matrix_wavedec(data_torch, wavelet, level=2)
print(coeff)
# backward 
rec, ifwt_matrix = ptwt.matrix_waverec(coeff, wavelet, level=2)
print(rec)