Open senpai-a opened 4 years ago
Hi @senpai-a
There are several reasons that might explain why the coefficients are shifted:
PDWT is tested against (somewhat old versions of) Pywavelets, and should be compatible with it, and therefore with the Matlab "wavelets" toolbox. I don't know how the RICE wavelets toolkit does the computations.
I see. Thx for the explaination. Closing this issue. BTW, is there a way I could cite this work? I might use it for some research.
Hello @pierrepaleo , can you generate a DOI for this repo so that I can cite.
Hello there. I'm using Wavelets::forward() to perfrom DWT and Wavelets::get_coeff() to get my result.
Here are 2 images I tested:
The wavelet used here is "db8", and here are the results:
I compared them with the output from another DWT library (Rice Wavelets Toolbox), who gives these:
I found that the spatial position of the coefficients seems to be shifted by a small amount.
I'm no an expert at signal processing so I don't know whether this is a bug or a non-critical issue.