Open QuasiLegendre opened 2 years ago
Hi @QuasiLegendre, it's not currently available in this repo but it wouldn't be too hard to add if it would be of use. As for the advantages of the DTCWT over the DWT - yes there are indeed! I'll refer you to the journal paper here: https://eeweb.engineering.nyu.edu/iselesni/pubs/CWT_Tutorial.pdf, in particular note the 4 problems mentioned in the introduction.
However, the DTCWT does add some complexity too, so it's hard to say if it's worth the extra effort. What's the use case you're thinking of?
Thanks for your advice! My use case is to find a optimized wavelet filtering/smoothing method for my particular 1-D real signal for signal anomaly detection in the next step(LSTM/GRU or 1-D CNN).
The signal is shown above, the red-labeled part is abnormal signal, while the blue part is referred as normal signal. The x axis means time, and y axis is voltage.
If you want to get good localisation of the signal then yes perhaps the DTCWT will be a good one to use. I can make a 1D version of it for you but it may take a couple of days. In the meantime, I'd suggest first trying with the DWT to see how well it works though.
I will take your advice and try DWT in several days. Thank you the recommendation and help!
@fbcotter any plans to release 1.3 with 1D DWT?
Do you mean a 1D DTCWT? There already should be a 1D DWT in there. I see I haven't made docs for it though, so maybe you just didn't see it.
from pytorch_wavelets import DWT1DForward, DWT1DInverse
@lostmsu no, I mean PyPi only has 1.2, so to get 1.3 one currently needs to install from source.
Oh I see. Sure, I can do that
Hey, did you ever get around to implementing the 1D DTCWT?
Hello, I would like to ask if you have implemented DTCWT to 1-D signals? Thank you so much.
Thanks for such a convenient framework! Is there anyway or method to implement DTCWT to 1-D signals? Also, does DTCWT has advantage over DWT in signal processing?