OCR extraction using Azure Vision

Athento allows you to extract OCR from a document using Azure Vision through automated OCR extraction.

Preliminary configuration

To extract OCR from a document using Azure Vision, we must:

• Create a Computer Vision resource in Azure.
• Configure two settings in Athento:
  • AZURE_VISION_ENDPOINT: Service URL. We will obtain this when creating the resource.
  • AZURE_VISION_KEY: Subscription key. We will obtain this when creating the resource.

OCR extraction

To perform OCR extraction using Azure Vision, we must use the “op_extract_ocr_v3” operation. In the operation parameters, we must select the “Azure vision” engine.

If the document is a PDF, we must check the “Force” parameter so that the OCR extraction is performed using Azure Vision.

Once configured, we can now run the automation on any document to extract the OCR.

**Note: If the “Azure Vision” option does not appear in the engine selector, reinstall the operation.