Addition of OCR and automatic generated name in note scanner

[peerzadaaabid]

Here I want to improve the note scanner app by incorporating the following two ideas

1. When we scan documents, we need to rename them individually, So I want to improve the by including the automatic generated name. The format of the name may be chosen like this 'Document Number date time' 2..OCR (OPTICAL CHARACTER RECOGNITION)

[allgood]

You can do this and submit a pull request, I can merge it after review.

On the document name, it is nice that the date and time be in YYYYMMDD-HHMM format right at the start of the file name, so sorting files will give the documents in the correct order.

About the OCR, please take a look on how to do it. The application should not make use of any external services and all libraries included should be compatible with the GPL version 3 license.

Best regards and thank you for your interest in helping ONS.

[bartatgithub]

Yes, OCR! I have installed this app because it looks that it does OCR, after a lot of tests and research I found out it didn't do OCR. ;-) Do you know this one: https://f-droid.org/en/packages/org.atai.TessUI/

But your app is nicer, so if you could implement the Tesseract library?

[allgood]

Great, will take a look at it! OCR was never the intent of Open Note Scanner, but it may be combined with other apps.

Em ter., 5 de nov. de 2019 às 10:58, bart notifications@github.com escreveu:

Yes, OCR! I have installed this app because it looks that it does OCR, after a lot of tests and research I found out it didn't do OCR. ;-) Do you know this one: https://f-droid.org/en/packages/org.atai.TessUI/

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ctodobom/OpenNoteScanner/issues/160?email_source=notifications&email_token=ABOKDUDDZWMMVX3ZISESTWLQSFUYTA5CNFSM4G5UV4S2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEDCXP3I#issuecomment-549812205, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABOKDUCH7KTXLZ4FYVL5EU3QSFUYTANCNFSM4G5UV4SQ .

[kollokollo]

I have a further idea: Once OCR is included (and faily working) it would be great if the App can also extract key information from the documents. Assume one scans an invoice or bill, one need to automatically extract the date, invoice number, owner, tax number, bank account etc... I know that this is not trivial, because there is no standardized format, invoices and bills are made. But getting the data out of it in an automatic way would improve all bookkeeping tasks. The extracted data could be included in the pdf as internal data structures, as a dataset encoded into a qr-code, or as part of the file name (customizable).

BTW: I also like the color mode, to produce artwork from photographs! One error encountered: I had an old version of opencvmanager installed and ONS crashed on start before allowing to download a newer version or even send a crash report to the developpers.

[rauferd]

This is probably off-topic and totall unrelated to the development of ONS (feel free to delete):

On my android, I use OpenNoteScanner together with this OCR app: https://github.com/SubhamTyagi/android-ocr In the OCR app, I choose ONS as image source. This way, you get searchable text from photos. Easy.

For automatic document processing, I recommend one of these projects:

• paperless-ng: https://github.com/jonaswinkler/paperless-ng
• Paperwork: https://openpaper.work/

Unfortunately, both of them are not intended to run on android devices. But if you have a spare raspberry pi or some hosted server available, they will probably serve your purpose well. They can auto-ocr, store, and tag your files and extract text and even run custom commands based on the extracted text to some extent.

P.S.: Many more projects like this exist, like Mayan EDMS, docspell etc. plus several closed source options. If anyone tested one of those with ONS, and would like to recommend something else, I'd love to hear about that.