The Model Library includes many different document extraction models that can be used for extracting information from your documents.
The models fall into two different categories, Predefined Models for common use cases, as well as Custom Models that can be customized for your exact requirements.
Predefined Models
These models have been pre-trained on millions of documents and images, providing very high accuracy out of the box in real time. They are designed to make getting started building document extraction as easy as possible.
The predefined models available are:
- US Driver's License OCR: Extracts information from Driver's License including names, dates, addresses and ID Numbers
- Passport OCR : Extracts information from Passports including names, dates and ID Numbers
- Receipt OCR : Extracts information from Receipts including Merchant Information, Transaction Amounts and Line Items
- Invoice OCR: Extracts information from Invoices including Vendor Information, Customer Information, Invoice Amounts and Line Items
- ID Card OCR: Extract information from ID Cards including Driver's License, Passports and State ID Cards
Custom Models
Custom models enable you to build and customize your own document extraction model for your exact use case. They are designed to give you full flexibility about what you'd like extracted out of your documents.
For details on how to create custom models you can check out this guide
For details about how to process documents using these predefined models, see the Testing Document Extraction and Using the REST API guide.