2D Autocorrelation Wavelets Transform

The 2D AC Wavelet transform will take a 2D signal such as an image and decompose it into wavelet coefficients. First, the row-wise coefficients will be computed using the 1D transform and then the column wise coefficients for each set of row-wise coefficients will be computed to create a 4 dimensional tensor.

Forward AC Wavelet Transform

There are 2 functions available for the 2D decomposition. The ac2d function is the original implementation and is kept for replication purposes. We generally recommend using the acwt2D function, which is a wrapper around ac2d because it has a better syntax.

Inverse AC Wavelet Transform

To reconstruct a 2D signal from a 4 dimensional tensor of wavelet coefficients, use the acwt2D function.