stegosphere.transform package

Submodules

stegosphere.transform.IWT module

Integer Haar Wavelet Transform

stegosphere.transform.IWT.inverse(coeffs, boundary_info)

Inverse transform data from wavelet domain into spatial domain using teh Integer Haar Wavelet Transform.

Parameters:

  • coeffs – Dictionary with wavelet coefficients from the transform function

  • boundary_info – Dictionary containing the boundary handling information as returned by the transform function

Returns:

Reconstructed numpy array

stegosphere.transform.IWT.transform(array, skip_last_axis=False, boundary_mode='discard')

Transform data into wavelet domain using the Integer Haar Wavelet Transform. Uneven dimensions are handled using skip_last_axis and boundary mode.

Parameters:

  • array – input numpy array

  • skip_last_axis – If True, the transformation is not applied along the last axis. Recommended for colour images.

  • boundary_mode – Method to handle uneven dimensions along axes being transformed. Options: - ‘discard’ : remove the last slice and restore later - ‘symmetric’/’reflect’/’edge’/’wrap’: use np.pad to make dimensions even, and remove padding after inverse transform

Returns:

A tupling containing: - dictionary with wavelet coefficients - information needed for the inverse transform

Module contents

The transform package currently only holds the Integer Wavelet Transform for fully reversible transformation.

(This is not to be confused with embeddings methods that rely on transforms).