Question

How to create an app to convert an OCR image to normal text?

1

Hello Technocrats, There is a challenge assigned for me to make an app to convert an OCR image to normal text .In these case i tried the following ideas but ended up reading the normal text but no numerical values , and also colored numerical values (i) Tesseract OCR module to read the values but ended up in the reading the text characters. (ii)Google OCR referred by youtube is not available in mendix marketplace? (iii) Google vision API How to implement Google vision API in mendix to get the OCR image to normal text Or any other method to implement for these app Thank you in advance.

asked 2023-10-16

THARUN BALAJI J S

8 answers

Tim van Steenbergen · Answer 1 · 2023-10-16

There are some OCR-services in the marketplace, one of them is mine, so that is of course the best one:-)

OCR-service

But also just go to the marketplace and search on OCR.

Bright A · Answer 2 · 2023-11-08

Hi THARUN BALAJI J S

You can use this Google Cloud Vision module

This module will helps you.

https://marketplace.mendix.com/link/component/208921.

THARUN BALAJI J S · Answer 3 · 2023-10-17

We are going over the google cloud vision API to get the REST api response from the post request can anyone give idea regarding this how to implement google cloud vision api in mendix

THARUN BALAJI J S · Answer 4 · 2023-10-17

I m finally using the google vision API to Mendix I have followed the procedure API key creation set the key value in resource folder .But the API response is 403 for all the image search. Is any roles and features to be enabled for the cloud settings?

Sjors Schultz · Answer 5 · 2023-10-18

I think you should probably set the http headers to use your API key instead of putting it in the resource folder. Have a look at Call REST Service | Mendix Documentation.

Flora Olivia · Answer 6 · 2023-10-19

After create, let me know. I am ready to share your app on my store.

THARUN BALAJI J S · Answer 7 · 2023-11-25

In this after enabling the billing account following the GOOGLE cloud vision module and their API recognize the image properly

Miroslav Samoylenko · Answer 8 · 2024-05-31

If you want to go fancy, try OpenAI module and rely on OpenAI Vision recognition. I just tested an example from Tim van Steenbergen module and ran it via copilot.microsoft.com (at the time of writing, this should be GPT 3.5 Turbo). It not inly recognized the text for Fenipic Plus, but also translated it to English for me.