Hi Oslien,
There are various options and uses for Optical Character Recognition (OCR) in Blue Prism; however, depending on your particular use case, some are better suited than others.
Using OCR in the automated application
This is the simplest form of OCR that Blue Prism supports. When conventional methods fail to capture (read) text from the application being automated then Surface Automation techniques can be employed. Part of Surface Automation is being able to recognise text displayed by the application allowing Blue Prism to capture the information. Blue Prism makes use of Google's Tesseract open source OCR engine to perform this function.
For more information please see Guide to Reading Text with OCR.
Using OCR to process data from other sources
Broadly speaking data which is being analysed by OCR can be split in to two categories: Structured and Non-Structured.
Non-structured data
This is the process of detecting and transcribing all recognised text in the source data without applying any formatting or structure to the output. Typically the output would be a simple string of text appearing in the same order that the OCR engine recognised it. Non-structured OCR is typically used to detect text within images, not for obtaining data from formal documents (such as invoices or forms).
To capture text from non-structured documents or images, the following three skills are provided on the Digital Exchange:
All of these options are cloud based AI powered machine learning OCR engines. To use these services an account must be registered with the service provider. Most of the service providers offer a limited free service, with paid subscriptions once the number of free API requests have been exceeded.
For additional information about these OCR services, please check out the links below:
------------------------------
InJoe Khor
Blue Prism
Blue Prism
------------------------------
In Joe Khor
Sr. Product Consultant
Blue Prism
