Search code examples
pythontensorflowautoencoder

Autoencoder Decoded Output

I am trying to build an AutoEncoder, where I am trying to de-noise the signal.

Now, for example, the amplitude range of my input signal varies in between -47 to +32. But while I am getting the decoded signal (reconstructed one), that signal only ranges in between 0 to +1 amplitude.

How can I get my reconstructed signal with the same amplitude range of -47 to +32?

Solution

  • I guess the question is whether your returned signal is a faithful representation of the input signal (but it's just constrained to the range 0 to 1)?

    If so, you could simply multiply it by 79, and then subtract 47.

    We'd need to see code if it's more than just a scaling issue.