Azure Translator Service is not able to translate Images in few PDFs

Question

Azure Translator Service is not able to translate Images in few PDFs

Rama Krishna 0

Azure Translator Service is not able to translate Images in few files

(PDF with Text) --> Translation is working

(Images downloaded and Attached to word and converted to PDF) --> Translation is working

(Snipped Image and Text or PDF merged with Scanned Images) --> Translation is not working

(Snipped Image only in PDF) --> Translation Working

(Complete Scanned PDF) --> Translation Working

santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-03T10:51:37.2533333+00:00

Hi @Rama Krishna,

Thank you for reaching out to Microsoft Q&A forum!

Azure Translator Service successfully translates documents in various file formats. PDF documents generated from digital file formats (also known as "native" PDFs) provide optimal output.

Translating scanned PDF files (Snipped Image and Text or PDF merged with Scanned Images) can result in loss of the original formatting, layout, and style, and affect the quality of the translation.

To improve translation accuracy with Azure Translator for PDFs merged with snipped images and text, start by enhancing the quality of snipped images to ensure they are clear and high-quality for OCR detection.

I hope you understand! Thank you.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-04T10:51:02.16+00:00

Hi @Rama Krishna,

Following up to see if the response was helpful. Thank you.
Rama Krishna 0 Reputation points

2024-07-04T11:02:09.8066667+00:00

Hi @santoshkc ,

Basically we have some good amount of PDF's which were Merged with "Normal PDF's" and "Scanned PDFs". In this case the content inside Normal PDF is translating but not the scanned. So Scanned PDF's has no issue with quality. we need to translate the content inside scanned pdf's. i will provide examples if required.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-04T11:29:36.02+00:00

Hi @Rama Krishna,

Thank you for you follow-up query.

To translate the content inside scanned PDFs, you'll need to perform OCR (Optical Character Recognition) to convert the scanned images into editable and translatable text. Azure Services offers an OCR API that can be used for this purpose. Once the OCR process is completed, you can then use the Azure Translator service to translate the extracted text. This approach ensures that the high quality of your scanned PDFs is maintained while enabling the translation of their content. Here's a suggested workflow: first, apply the OCR API to extract text from the scanned PDFs, and then feed the extracted text into the Azure Translator service to achieve your translation objectives.
See: Document Intelligence Read OCR model

I hope you understand. Thank you.
Rama Krishna 0 Reputation points

2024-07-04T12:40:16.69+00:00

Hi @santoshkc ,

Thank you for your prompt response. But this will return a json right. But our requirement is Complete Document translation.(Translating Foreign language PDF document to English Language PDF).
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-05T05:47:28.16+00:00

Hi @Rama Krishna,

Yes, it will return a format of json. I request you to raise a support request through Azure portal. They will be able to investigate the issue further and provide a more targeted solution.

Thank you.

Your answer

santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-03T10:51:37.2533333+00:00

Hi @Rama Krishna,

Thank you for reaching out to Microsoft Q&A forum!

Azure Translator Service successfully translates documents in various file formats. PDF documents generated from digital file formats (also known as "native" PDFs) provide optimal output.

Translating scanned PDF files (Snipped Image and Text or PDF merged with Scanned Images) can result in loss of the original formatting, layout, and style, and affect the quality of the translation.

To improve translation accuracy with Azure Translator for PDFs merged with snipped images and text, start by enhancing the quality of snipped images to ensure they are clear and high-quality for OCR detection.

I hope you understand! Thank you.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-04T10:51:02.16+00:00

Hi @Rama Krishna,

Following up to see if the response was helpful. Thank you.
Rama Krishna 0 Reputation points

2024-07-04T11:02:09.8066667+00:00

Hi @santoshkc ,

Basically we have some good amount of PDF's which were Merged with "Normal PDF's" and "Scanned PDFs". In this case the content inside Normal PDF is translating but not the scanned. So Scanned PDF's has no issue with quality. we need to translate the content inside scanned pdf's. i will provide examples if required.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-04T11:29:36.02+00:00

Hi @Rama Krishna,

Thank you for you follow-up query.

To translate the content inside scanned PDFs, you'll need to perform OCR (Optical Character Recognition) to convert the scanned images into editable and translatable text. Azure Services offers an OCR API that can be used for this purpose. Once the OCR process is completed, you can then use the Azure Translator service to translate the extracted text. This approach ensures that the high quality of your scanned PDFs is maintained while enabling the translation of their content. Here's a suggested workflow: first, apply the OCR API to extract text from the scanned PDFs, and then feed the extracted text into the Azure Translator service to achieve your translation objectives.
See: Document Intelligence Read OCR model

I hope you understand. Thank you.
Rama Krishna 0 Reputation points

2024-07-04T12:40:16.69+00:00

Hi @santoshkc ,

Thank you for your prompt response. But this will return a json right. But our requirement is Complete Document translation.(Translating Foreign language PDF document to English Language PDF).
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2024-07-05T05:47:28.16+00:00

Hi @Rama Krishna,

Yes, it will return a format of json. I request you to raise a support request through Azure portal. They will be able to investigate the issue further and provide a more targeted solution.

Thank you.

Share via

Azure Translator Service is not able to translate Images in few PDFs

Your answer