ABBYY FineReader Engine 11 for Windows - Release 2
Release date: 06 March 2014 (public release)
Part #: 1041/20
Build #: 126.96.36.199
- Ability to inject the text layer into an existing PDF - This new method creates a copy of the input PDF file and adds a new text layer that corresponds to the recognized text of the document.
- Support of PDF/A-3a and PDF/A-3u
- Ability to write attachments from the original PDF file into the output file. (Ignored if the original file was not a PDF).
- New Java code samples
- BatchProcessing - shows how to use Batch Processor for processing a large amount of one-page documents.
- EventsHandling - illustrates the use of the callback interfaces using the FRDocument callback interface as an example. The sample shows the progress of recognition and export during image processing. The callback interfaces can be used to control image processing
- EnginesPool - provides a complete reusable solution for a pool of Engines in a multithreaded application. This sample can be used to observe the speed increase when using multiprocessing.
The Java samples support the same scenarios as the samples with the corresponding title provided in C#.
- Full version of Java wrapper - the complete version of JAVA wrapper is implemented, allowing to access the entire SDK functionality through the Java wrapper.
Enhancements in recognition
Ability to detect vertical text allows detecting vertical text in all languages (not only CJK).
A new predefined profile for recognition of engineering drawings and maps allows extractin of text in various types of orientation and conversion of engineering drawings from PDF to searchable PDF.
New predefined languages OcrA and OcrB - the OCR A and OCR B can now be selected from the list of predefined languages (in addition to the OCR A and OCR B text types). They contain regular characters as well as special symbols from the OCR A and OCR B standards.
New special predefined languages:
English_US_Address_PhoneNumber (contains phone numbers (English, United States).
English_US_WellKnownCode_EIN (contains Employer Identification Numbers (English, United States).
- Support for “IntelligentMail” / USPS 4-CB barcodes
IntelligentMail / USPS 4-CB barcodes are now supported. Intelligent Mail is a height-modulated 65-bar barcode which is used on mail in the United States. BT_ IntelligentMail was added to BarcodeTypeEnum enumeration constants.
- Improved barcode detection - a new property allows detecing more barcodes (but can negativelly impact speed)
- Ability to define the minimal height in relation to average letter size - allows to better detect small barcodes.
Enhanced screenshot detection
FineReader Engine 11 is able to identify screenshots in documents and keep them as images (without extracting the text)
Enhancments in classification
New method for classification training allows adding an already recognized page to the classification database.
Ability to add custom features for classification - new methods were added supporting training the classification database with the help of user-defined additional features that allow increasing the accuracy of classification.
Export with original layout to XLSX
A new property allows to export files to XLSX with original layout and to choose the mode of retaining tables’ formatting during export to XLSX.
Enhanced JBIG2 lossless compression
The technology for lossless compression for black and white JBIG2 images was significantly improved.
Improved key value tables detection
Key value tables’ detection was significantly improved.
Improved memory allocation
Memory fragmentation was significantly reduced to avoid out of memory issues.
Please sign in to leave a comment.