AI-multimodal / aimm-post-processing

This repository is depricated. We will be doing all postprocessing in `aimmdb.postprocessing`.
Other
0 stars 3 forks source link

Implement filtering #11

Open matthewcarbone opened 2 years ago

matthewcarbone commented 2 years ago

Zhu's comment:

Filtering: this step essentially filters out “bad” spectra according to, and not limited to, the following criteria: Noise. Filters out noise spectra (some noisy spectra might still be useful, but we need to talk to Eli after implementing this. Or an “averaging” step is also necessary). Shape. Those spectra with some weird shape will be filtered out (this step is by far done by some human defined rule. A fully-connected network was also created to do the job, but I stopped improving it at some point, need to revisit it).

@zhul9311, can you elaborate a bit more on what exactly you mean here? Specifically.

  1. By filtering out bad spectra, do you mean just letting the user know that it's "bad" by some user-defined criteria?
  2. By filtering out spectra with weird shapes, what exactly do you mean? Can you be specific? Makes it easier to code it up.

Note as well, I don't know if this will fall under the category of the Operator class, but it might (perhaps it just returns None if the spectrum is bad, along with the appropriate metadata).