From the course: Microsoft Azure AI Essentials: Workloads and Machine Learning on Azure

Azure AI Document Intelligence demo

From the course: Microsoft Azure AI Essentials: Workloads and Machine Learning on Azure

Azure AI Document Intelligence demo

“

- [Instructor] The capabilities of Azure AI document intelligence can be demonstrated using the Document Intelligence Studio. To get started, you'll need a document intelligence resource. In the Azure portal, search for document intelligence, click Create, and assign the correct resource group resource name, location, and pricing tier. Click Review plus Create, then click Create. Once the resource is ready, search Azure Document Intelligence Studio in Bing to find the studio. Log in with your Azure credentials and click the settings cog icon to verify you're using the correct resource. You'll see various documents like document analysis, prebuilt models, and the option to create custom models. Let's explore the layout model. You can choose from several samples. Click Run Analysis to see how the system captures elements like page header, title, section headings, paragraph, and many more. The Result tab lets developers extract data in JSON format, while the Code tab provides sample code for Python, JavaScript, or C#. Running analysis on other documents shows the system's ability to handle diverse formats, including scanned documents. The layout checklist sample, for instance, demonstrates its capacity to capture scanned images. While Azure AI Vision captures texts in scanned images, it doesn't recognize document structure, which makes Azure AI document intelligence crucial for more detailed extraction. For example, the Receipts model identifies key information from retail receipts, capturing line items, merchant details, subtotal, tip, total, and so on. Running the analysis on other Receipt formats proves its versatility, including the ability to process scanned receipts. These prebuilt models are powerful tools for extracting structured data from both digital and scanned documents.

Contents