Use prebuilt Document intelligence models

Unlocking the Power of Prebuilt Models with Azure AI Document Intelligence

In today’s fast-paced business environment, efficient document processing is crucial. Azure AI Document Intelligence offers a suite of prebuilt models designed to streamline the extraction of data from common forms and documents, eliminating the need for custom model training. This blog explores the capabilities and features of these prebuilt models, demonstrating how they can enhance your document processing workflows.

What Are Prebuilt Models?

Prebuilt models in Azure AI Document Intelligence are pre-trained on large datasets of specific document types, allowing you to extract data without the need for custom training. These models are designed to handle common business documents such as invoices, receipts, tax forms, and more. By leveraging these models, businesses can achieve accurate and reliable data extraction with minimal effort.

Key Prebuilt Models

1. Invoice Model: Extracts fields like customer name, invoice total, and due dates from invoices, even if they are poorly scanned or at odd angles.
2. Receipt Model: Identifies merchant details, transaction amounts, and itemized purchases from receipts.
3. US Tax Model: Handles forms such as W-2, 1099, and 1040, extracting relevant tax information.
4. ID Document Model: Extracts data from US driver’s licenses, EU IDs, and international passports.
5. Business Card Model: Captures contact information from business cards, including names, addresses, and phone numbers.
6. Health Insurance Card Model: Extracts information from health insurance cards.
7. Bank Statement Model: Retrieves account information and transaction details from bank statements.
8. Pay Stub Model: Extracts wages, hours, deductions, and net pay from pay stubs.

Features of Prebuilt Models

Prebuilt models offer a range of features to ensure comprehensive data extraction:

• Text Extraction: Captures printed and handwritten text.
• Key-Value Pairs: Identifies labels and their corresponding values.
• Entities: Recognizes complex data structures such as people, locations, and dates.
• Selection Marks: Detects choices indicated by radio buttons and checkboxes.
• Tables: Extracts data from tables, including cell content and structure.

Input Requirements

To achieve optimal results, ensure your documents meet the following criteria:

• Formats: JPEG, PNG, BMP, TIFF, or PDF.
• Size: Less than 500 MB for standard tier, 4 MB for free tier.
• Dimensions: Between 50 x 50 pixels and 10,000 x 10,000 pixels.
• PDF: Less than 17 x 17 inches or A3 size, not password-protected.

Using Azure AI Document Intelligence Studio

Azure AI Document Intelligence Studio provides a visual interface to experiment with prebuilt models. You can upload your documents or use sample documents provided by Microsoft to see how the models perform. This tool is invaluable for understanding model behavior and refining your document processing workflows.

Calling Prebuilt Models via APIs

Azure AI Document Intelligence supports RESTful web services, making it easy to integrate with your applications. APIs are available for various programming languages, including C#, Python, Java, and JavaScript. To use these APIs, you need the service endpoint and API key from your Azure subscription.

Conclusion

Azure AI Document Intelligence’s prebuilt models offer a powerful solution for automating document processing tasks. By leveraging these models, businesses can save time, reduce errors, and improve efficiency.

To learn more about creating intelligent document processing solutions with Azure AI Document Intelligence, join our free event on December 10, 2024: Create an Intelligent Document Processing Solution with Azure AI Document Intelligence.