Redaguoti

Bendrinti naudojant


Which model should I choose?

Important

  • Document Intelligence public preview releases provide early access to features that are in active development. Features, approaches, and processes may change, prior to General Availability (GA), based on user feedback.
  • The public preview version of Document Intelligence client libraries default to REST API version 2024-07-31-preview.
  • Public preview version 2024-07-31-preview is currently only available in the following Azure regions. Note that the custom generative (document field extraction) model in AI Studio is only available in North Central US region:
    • East US
    • West US2
    • West Europe
    • North Central US

Azure AI Document Intelligence supports a wide variety of models that enable you to add intelligent document processing to your applications and optimize your workflows. Selecting the right model is essential to ensure the success of your enterprise. In this article, we explore the available Document Intelligence models and provide guidance for how to choose the best solution for your projects.

The following decision charts highlight the features of each supported model to help you choose the model that best meets the needs and requirements of your application.

Important

Be sure to check the language support page for supported language text and field extraction by feature.

Pretrained document-analysis models

Document type Example Data to extract Your best solution
A generic document. A contract or letter. You want to primarily extract written or printed text lines, words, locations, and detected languages. Read OCR model
A document that includes structural information. A report or study. In addition to written or printed text, you need to extract structural information like tables, selection marks, paragraphs, titles, headings, and subheadings. Layout analysis model
A structured or semi-structured document that includes content formatted as fields (keys) and values. A form or document that is a standardized format commonly used in your business or industry like a credit application or survey. You want to extract fields and values including ones not covered by the scenario-specific prebuilt models without having to train a custom model. **Layout analysis model with the optional query string parameter features=keyValuePairs enabled **

Pretrained scenario-specific models

Document type Data to extract Your best solution
US Unified Tax You want to extract key information across all tax forms of W2, 1040, 1090, 1098 from a single file without running any custom classification of your own. US Unified tax model
US Tax W-2 tax You want to extract key information such as salary, wages, and taxes withheld. US tax W-2 model
US Tax 1098 You want to extract mortgage interest details such as principal, points, and tax. US tax 1098 model
US Tax 1098-E You want to extract student loan interest details such as lender and interest amount. US tax 1098-E model
US Tax 1098T You want to extract qualified tuition details such as scholarship adjustments, student status, and lender information. US tax 1098-T model
US Tax 1099(Variations) You want to extract information from 1099 forms and its variations (A, B, C, CAP, DIV, G, H, INT, K, LS, LTC, MISC, NEC, OID, PATR, Q, QA, R, S, SA, SB). US tax 1099 model
US Tax 1040(Variations) You want to extract information from 1040 forms and its variations (Schedule 1, Schedule 2, Schedule 3, Schedule 8812, Schedule A, Schedule B, Schedule C, Schedule D, Schedule E, Schedule EIC, Schedule F, Schedule H, Schedule J, Schedule R, Schedule SE, Schedule Senior). US tax 1040 model
Bank Statement You want to extract key information from US bank statement \Bank Statement
Bank check You want to extract key information from check document. Bank Check
Contract (legal agreement between parties). You want to extract contract agreement details such as parties, dates, and intervals. Contract model
Health insurance card or health insurance ID. You want to extract key information such as insurer, member ID, prescription coverage, and group number. Health insurance card model
Credit/Debit card You want to extract key information bank cards such as card number and bank name. Credit/Debit card model
Marriage Certificate You want to extract key information from marriage certificates. Marriage certificate model
Invoice or billing statement You want to extract key information such as customer name, billing address, and amount due. Invoice model
Receipt, voucher, or single-page hotel receipt. You want to extract key information such as merchant name, transaction date, and transaction total. Receipt model
Identity document (ID) like a U.S. driver's license or international passport You want to extract key information such as first name, surname, date of birth, address, and signature. Identity document (ID) model
Pay stub You want to extract key information from the pay stub document. Pay stub Model
US Mortgage 1003 You want to extract key information from the Uniform Residential loan application. 1003 form model
US Mortgage 1004 You want to extract key information from the Uniform Residential Appraisal Report (URAR). 1004 form model
US Mortgage 1005 You want to extract key information from the Verification of employment form 1005 form model
US Mortgage 1008 You want to extract key information from the Uniform Underwriting and Transmittal summary. 1008 form model
US Mortgage Closing Disclosure You want to extract key information from a mortgage closing disclosure form. Mortgage closing disclosure form model
Mixed-type document(s) with structured, semi-structured, and/or unstructured elements You want to extract key-value pairs, selection marks, tables, signature fields, and selected regions not extracted by prebuilt or general document models. Custom model

Tip

  • If you're still unsure which pretrained model to use, try the layout model with the optional query string parameter features=keyValuePairs enabled.
  • The layout model is powered by the Read OCR engine to detect pages, tables, styles, text, lines, words, locations, and languages.

Custom extraction models

Training set Example documents Your best solution
Structured, consistent, documents with a static layout. Structured forms such as questionnaires or applications. Custom template model
Structured and semi-structured. ● Structured → surveys
● Semi-structured → invoices
Custom neural model
Unstructured documents, documents with varying templates. ● Unstructured documents like contracts or letters
● Varying document templates like loan statements from different mortgage companies
Custom generative model
A collection of several models each trained on similar-type documents. ● Supply purchase orders
● Equipment purchase orders
● Furniture purchase orders
All composed into a single model.
Composed custom model

Custom classification model

Training set Example documents Your best solution
At least two different types of documents. Forms, letters, or documents Custom classification model

Next steps