1

In the application of the Huang Hilbert Transform (HHT) (or aka EMD, Empirical Mode Decomposition), there is the notable challenge of handling the edges (boundaries) of the signals; as noted in the answer link to its mention in other question.

What are some practical techniques that can be applied to signals in which there is 'noise' padding already present and when there is no redundant signal information at the boundaries? (eg how would a mirroring strategy be applied?)

Vass
  • 125
  • 7

1 Answers1

1

One possibility is to mirror the extrema near both edges with respect to the edge time values. I think this is one of the best ways to reduce edge effects. This paper explains better the way to do it.

https://www.sciencedirect.com/science/article/abs/pii/S0888327007002701

Hope this is helpful to you.

EmThorns
  • 393
  • 3
  • 11