I have managed to get my SCIO to identify recyclable plastics. Luckily this was easy to do as most recyclable plastics carry recycling symbols identifying what they are made of. The test model facility of scio lab needs a feature to allow me to mark the answer from the model as right or wrong so that I can verify my various models accuracies. I have made multiple models for the same sample set and I need to score each model to see which is the best model. Also by scoring the models you could use genetic algorithms to literally evolve scio models.
I would also like to show a calculated percentage match of the spectra to the nearest matching model. i.e If I point my SCIO at an unknown substance I would like to report a no match condition if the certainty of the match fell below a preset percentage level, rather than reporting a match.
Collections of samples should be loose so that a sample can belong to two or more collections. So for example with my plastics model I have two attributes colour and plastic type. I can use my eyes to determine the colour of the plastic so I could define multiple models based on the colour attribute, using standard set theory. The plastics collection might belong to a super set of all polymers, which in turn belongs to the universal set off all known materials. With this scheme I could make a model of all blue polymers for example, which would include the subset of all blue plastics.
I would also like the ability to be able to subtract a base spectra from a sample spectra to help aid with the identification of a mystery substance if this is possible.
Mixture[Solvent Spectra + Solute Spectra] – Pure[Solvent Spectra] = Pure[Solute Spectra] if this is possible ?
I would also like the ability to allow collaborators to upload scio scans to my collections from there own scio’s. One of the applications I have in mind would require lab colleagues to identify a mystery substance and then use the scio to create a spectra of it for upload.