Azure AI Document Intelligence recognition engine
This is a cloud-based recognition engine that uses Azure AI Document Intelligence. This is available for both page and zone recognition.
This recognition engine is composed of multiple advanced machine-learning based models that support global languages. It can extract printed and handwritten text including mixed languages and writing styles.
In order to use the Azure AI Document Intelligence for page or zone recognition, you require either a Tungsten provided API key or a Microsoft Azure account and an API key for access. This is configured in the Project Settings - Service Endpoints tab.
When configuring a page or zone profile, there is no way of specifying a language. This is because Azure AI Document Intelligence is able to successfully determine the language of a document on its own. It is also able to cope with documents that contain content in more than one language, even on a single line of text. A user interface setting could hinder the recognition engine in any situations where the language of a document differs from the language specified by the setting.
This recognition engine is not available for a mixed print recognition profile.
Licensing of the Azure AI document intelligence service
In order to use this recognition engine, a license is required. There are two different ways that you can license this recognition engine.
-
Use a Tungsten provided volume license.
-
Use your own endpoint and API key provided by Microsoft.
A Tungsten provided volume license charges for production use only, whereas you are charged by Microsoft each time your own API key is used, including within Quick Capture or the Transformation Designer.
The advantage of a Tungsten provided volume license is that you can configure and test a project in the Transformation Designer or a solution in Quick Capture without decrementing the volume license.
Related topics: