Medicine recognition

[viraj-dhanushka]

This PR is relevant to the issue #26

Here I used firebase_ml_vision plugin.

This plugin uses the capabilities of Firebase ML, which includes all of Firebase's cloud-based ML features, and ML Kit, a standalone library for on-device ML, which can be used with or without Firebase.

NOTE: The below functionality is recorded in a non internet connected environment. So this approach seems to be very suitable for the considering scenario.

Using the live camera

Image Used

Uploading the Image

Image Used

[im-Amitto]

@viradhanus we can't use these as test data, can you please try it on an actual medicine box and using a camera instead of gallery.

[viraj-dhanushka]

@im-Amitto actually its working just fine and I did test for few. One is as follows.

[im-Amitto]

So what database it use to extract medicine name from random text?

[viraj-dhanushka]

Here only the text recognition part has implemented. The specialty here is, it is an standalone library which can work just fine without a internet connection. The extracting part has to be implemented.

[im-Amitto]

As long as I remember, till now we had 2 text recognition implemented.

1. OCR which needed the go lang server running
2. Firebase ml which worked offline as well as online (gives better results when online)

Work which I was expecting: Use any existing option to gain meaningful data out of the extracted text.

[viraj-dhanushka]

@im-Amitto Sure I will try with Firebase ML and update this PR

[im-Amitto]

I think what you did here is working better than what we already have so if you want you can go ahead with this as well but our main goal is to get meaningful data out of the scanned text so feel free to continue in this PR to make it work.

Also, please find and delete any existing code which is not needed when you add it's counter part. We don't need zombie code in or codebase.

[viraj-dhanushka]

@im-Amitto Sure I will improve this with some drug name extracting technique and update this PR.

[im-Amitto]

@bradymiller shared a few good database when I was working on it but I lost that info. Sorry for that. Please have a talk with him, it will help you a lot.

[viraj-dhanushka]

@im-Amitto Thanks a lot for the info. I'll contact @bradymiller . :+1:

[bradymiller]

hi @viradhanus , Check out this api: https://rxnav.nlm.nih.gov/PrescribableAPIs.html#

[viraj-dhanushka]

Thanks a lot @bradymiller

[viraj-dhanushka]

Here I used DailyMed api to extract the drug names. A simple UI is added as follows.

[im-Amitto]

Thanks, @viradhanus Let's merge it and call it a kind of prototype.