OCR Gets an Upgrade – What’s New in the OmniPage Capture SDK

We come across new applications for Optical Character Recognition (OCR) technology in some of the unlikeliest of places. A recent example involves the use of computer vision in the online gaming industry. However, even more established applications like Enterprise Content Management (ECM) or Data Loss Prevention (DLP) are enjoying renewed innovation due to cutting-edge advances in the underlying OCR technology. 

For example, DLP services, Modern DLP applications promise to monitor, and keep safe, all employee communications that occur while using an organization’s IT resources. This includes computers, mobile devices, multifunction printers, fax machines and more. Until now, this has meant examining emails, text messages, faxes, printouts, copies and scanned documents. What has not been feasible is the ability to monitor what’s being viewed or typed on an employee’s computer screen. This is important for a number of reasons, including the ability to recognize a potential data breach before it happens – instead of just notifying security after the information has been sent and the breach has occurred. 

With the availability of the Kofax OmniPage Capture SDK toolkit version 21 for Windows, we deliver on the promise of improving upon the core technology that runs today's critical machine learning, DLP, RPA and document conversion solutions. This release expands upon the available programming languages developers can leverage to integrate OCR technology into their applications.

Key New Capabilities 

New enhancements to the toolkit make it more broadly accessible and easier to use. How? By doing more of the heavy-lifting for the developer. 

Improved OCR Accuracy 

The core benefit delivered by an OCR toolkit is the accurate conversion of text from an image. OmniPage Capture SDK implemented new algorithms for processing mixed alphanumeric strings, leading to accuracy improvement of up to 67%. Better accuracy means less exception handling for developers, and potentially less manual work for end users. 

Screen Capture OCR 

With the addition of the Screen Capture OCR enhancements, developers easily recognize data from screen capture images, allowing them to automate even more processes that previously required manual intervention. Screen capture images are in a special RGB format that previously could not be binarized well. This new capability features several special algorithms to convert this format into better binary images for higher OCR accuracy and is very useful to those developing AI/ML, DLP and RPA systems. 

Java Programming Interface 

Many big data infrastructure services, like Hadoop, utilize the Java programming language for developers that are interested in integrating with it. Those same developers now have a Java API to provide support for calling/executing OmniPage Capture SDK functions from an application written in the Java programming language.