How to extract the data from Prescription which is in Hindi language

Help AI Center
,
1

I want to extract the data from prescription. The prescription is in Hindi language.
The data’s I need to extract (Doctor Name, Mobile Number, Email, Address)

2

@Rudhresh_tamil,

If you are trying to extract letterhead details which are printed and not written by doctor, you can use OCR extraction or Document Understanding for it.

Thanks,
Ashok :slightly_smiling_face:

3

@Rudhresh_tamil

Hindi is under supported language for ocr …did you try to extract using du tempalte?

Cheers

4

@Anil_G and @ashokkarale Thanks for replying.

My requirement is needed to extract the data from prescription which is in JPG and JPEG format. Some data is in letterhead and sometime doctor name is in handwritten.

I used Document understanding in AI Centre. While labeling the document itself it cannot recognize the correct text from the image. For your reference I have attached the document below.

I used this ML package → out of box packages → UiPath Document understanding → UiPath Document understanding.

Pipeline Type - Train run

In UiPath studio used Omnipage OCR for digitize document.

Extractor - ML Extractor.

The image displays a document where text written in Hindi is incorrectly transcribed into Latin characters, along with a section showing a mobile number and a label indicating "Doctor Name" with the incorrectly transcribed text. (Captioned by AI)
The image displays a document where text written in Hindi is incorrectly transcribed into Latin characters, along with a section showing a mobile number and a label indicating "Doctor Name" with the incorrectly transcribed text. (Captioned by AI)822×315 85.9 KB

Please let me know what I need to correct

5

@Rudhresh_tamil

indocuemnt understanding project settings did you happen to choose the ocr like below

The image shows a dropdown menu under the "OCR method" section, where "UiPath Extended Languages OCR (Preview)" is highlighted among other OCR options. (Captioned by AI)
The image shows a dropdown menu under the "OCR method" section, where "UiPath Extended Languages OCR (Preview)" is highlighted among other OCR options. (Captioned by AI)1308×402 26.1 KB

cheers

6

Yes, I have selected the UiPath Extended language and OCR URL is auto generated but I cannot find the OCR API Key. Can you tell me where to find that.

7

@Rudhresh_tamil

if you see ocr key would be mentioned as optional and auto populated

just select apply ocr as yes

cheers

8

@Rudhresh_tamil

If yous elect everythign properly it should be working

I just tried and it is extracting

The image is a screenshot of a tax invoice document from a company named उज्जवल ट्रेडर, highlighting the "extractname" as उज्जवल, along with various details such as address, GST number, date, and client information. (Captioned by AI)
The image is a screenshot of a tax invoice document from a company named उज्जवल ट्रेडर, highlighting the "extractname" as उज्जवल, along with various details such as address, GST number, date, and client information. (Captioned by AI)1891×645 173 KB

steps

  1. Crate AI center project
  2. Upload docs in dataset
  3. Create a labeling session
  4. In labelign session first select the ocr engine as extended and select apply as yes
  5. then import docs and label

cheers

9

Thanks @Anil_G It’s working now for the newly uploaded file not for the existing file.

1 Like
10

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.