- Overview
- Getting started
- Activities
- Insights dashboards
- Document Understanding Process
- Quickstart tutorials
- Framework components
- ML packages
- Overview
- Document Understanding - ML package
- DocumentClassifier - ML package
- ML packages with OCR capabilities
- 1040 - ML package
- 1040 Schedule C - ML package
- 1040 Schedule D - ML package
- 1040 Schedule E - ML package
- 1040x - ML package
- 3949a - ML package
- 4506T - ML package
- 709 - ML package
- 941x - ML package
- 9465 - ML package
- ACORD131 - ML package
- ACORD140 - ML package
- ACORD25 - ML package
- Bank Statements - ML package
- Bills Of Lading - ML package
- Certificate of Incorporation - ML package
- Certificate of Origin - ML package
- Checks - ML package
- Children Product Certificate - ML package
- CMS 1500 - ML package
- EU Declaration of Conformity - ML package
- Financial Statements - ML package
- FM1003 - ML package
- I9 - ML package
- ID Cards - ML package
- Invoices - ML package
- Invoices Australia - ML package
- Invoices China - ML package
- Invoices Hebrew - ML package
- Invoices India - ML package
- Invoices Japan - ML package
- Invoices Shipping - ML package
- Packing Lists - ML package
- Payslips - ML package
- Passports - ML package
- Purchase Orders - ML package
- Receipts - ML Package
- Remittance Advices - ML package
- UB04 - ML package
- Utility Bills - ML package
- Vehicle Titles - ML package
- W2 - ML package
- W9 - ML package
- Other Out-of-the-box ML Packages
- Public endpoints
- Traffic limitations
- OCR Configuration
- Pipelines
- OCR services
- Supported languages
- Deep Learning
- Licensing
Document Understanding User Guide
Digitization overview
Digitization is the process of obtaining machine readable text from a given incoming file, so that a robot can then understand its contents and act upon them. It is the first step applied on files that need to be processed through the Document UnderstandingTM framework.
The digitization step has two outputs:
- the text from the processed file, stored in a string variable, and
- the Document Object Model of that file - JSON object containing basic information such as name, content type, text length, the number of pages, as well as detailed information such as page rotation, detected language, content and coordinates for every word identified in the file.
In the Document Processing Framework, digitization is performed using the Digitize Document activity.
Even though related, the digitization step is not OCR.
In many cases, the files that need to be processed are native PDF files (not scanned), that can be read programmatically by the robot without applying OCR.
The Digitize Document activity requires, as part of its configuration, the selection of an OCR engine - so that, at need, it can be used, but only executes OCR on:
- files that are images
- supported images formats are .png, .jpe, .jpg, .jpeg, .tiff, .tif, .bmp
- for multi-page TIFF files, OCR is applied for each page
- PDF pages that
- do not expose any machine readable content
- contain images that cover a significant area of the page.
- There is a 160 MB file size limit.
- There is a maximum 500 pages per document limit.
OCR is also applied, always, if the Digitize Document activity is configured with the ForceApplyOCR flag set to True. This option is usually recommended for use cases in which a significant percentage of files seem to contain native content, but the natively read content does not correspond to what a user can see in those files.
As each use case has its own particularities, it is strongly recommended to test all available OCR Engines with different settings, in order to determine which one works best for your project. Another recommendation is to pay particular attention to the OCR engine arguments, such as Profile, Scale, Language etc. (may vary from one engine to another), so that you identify the best settings for each use case.
