ABBYY Launches FineReader Engine 10 OCR SDK for Windows Systems
ABBYY, a leading provider of document recognition, document capture, and linguistic technologies and services, today announced ABBYY FineReader Engine 10 for Windows, the next evolution of its software development kit (SDK) to integrate cutting-edge document recognition technologies. FineReader Engine 10 can recognize documents in key European languages up to 92 percent faster than previous version*. The product also raises the bar for Asian language OCR (optical character recognition) with up to a 40 percent accuracy increase**, delivers extended image and photo processing capabilities, and an enhanced algorithm for document structure and formatting reconstruction. In addition, its new API based on an intuitive task-specific approach enables for maximum efficiency at minimum time and effort.
“The true challenge is not just to speed up OCR processing but, at the same time, to keep up the highest level of accuracy customers expect from OCR technology,” said Andrey Isaev, director of technology products department at ABBYY. “FineReader Engine 10 represents an ideal balance between those two basic criteria.”
The FineReader Engine SDK gives developers and integrators the tools they require to integrate a variety of recognition technologies into their applications. The ABBYY recognition platform delivers award-winning OCR, intelligent character recognition (ICR), barcode, checkmark, field-level/zonal recognition and PDF conversion for transforming scanned documents and images into searchable and editable text files. FineReader Engine 10 features the broadest language support covering 198 languages for OCR and 113 languages for ICR. The key new features and functions of ABBYY FineReader Engine 10 include:
Enhanced Fast Recognition Mode — Delivers up to 92 percent faster page throughput for European languages while maintaining the high level of recognition accuracy. It is designed to optimize processing speed/accuracy balance for good quality images.
Increased OCR Accuracy for Chinese, Japanese and Korean — Improved Asian character classifier providing a considerable OCR accuracy increase on documents in Chinese, Japanese and Korean, as well as, multilingual documents combining Asian and European characters.
New Recognition Mode for Low Resolution Scans — Delivers up to 20 percent accuracy improvement** on low resolution scans and low print quality documents (e.g. old faxes).
New Powerful API for Simpler Development — New API offers a set of pre-defined profiles for the most popular recognition tasks such as “document conversion for archiving,” “document conversion for content reuse,” “barcode recognition” and more. Profiles include optimal sets of processing parameters to start and the best OCR quality without the need for manual system tuning or a long learning curve. Manual tuning is available for specific custom solutions.
Improved Document Structure and Page Layout Reconstruction — Delivers further enhancements in the analysis of the logical structure of multipage documents via ABBYY’s second generation ADRT® (Adaptive Document Recognition Technology). FineReader Engine 10 accurately re-creates complex formatting attributes of original documents in native Microsoft® Office format — now including tables of contents, hierarchical heading structures, picture and table captions, and providing improved reproduction of footnotes and styles. The product also delivers enhanced preservation of document layouts containing background image items with text areas upon them and more accurately reproduces diagrams, graphs and charts in output files.
In addition, the SDK delivers a new Document Structure API, a unique feature-set to easily access and manipulate the document structure attributes such as headings, chapters, page numbers, footnotes, headers and more.
New Achievements in Image and Photo Processing — Enhanced ABBYY Camera OCR technology intelligently corrects image 3D distortions, ISO noise and motion blur. FineReader Engine 10 delivers new binarization technology that ensures entire text retention on the image, preventing the loss of information even in difficult cases. In addition, the solution includes a unique feature to separate color marks and stamps from the rest of the text on an image, improving recognition accuracy.
Improved PDF Export — FineReader Engine 10 offers predefined PDF export profiles — including MaxQuality, MinSize, Balanced, and MaxSpeed — allowing integrators to achieve better quality PDF conversion projects with a smaller file size.
Availability and Pricing
ABBYY FineReader Engine 10 for Windows is available immediately worldwide via a flexible, modular licensing policy. Developers may select the best combination of tools and pricing options for their projects. Pricing varies according to the number of CPU cores, processing stations and number of pages processed. For information on licensing models and pricing, contact your local ABBYY office or ABBYY partner. A time-limited trial version is also available for testing. For more information about the product, visit http://www.abbyy.com/ocr_sdk/.
* Based on internal testing of FineReader Engine 9, release 1, vs. FineReader Engine 10, release 1. Tests included standard business document scans in English, German, French, Spanish and Italian. (Your results may vary based on scan quality, document complexity, and system and application type.)
** Based on internal testing.