Software

ABBYY FlexiCapture 12: Recognition

The recognition stage of ABBYY FlexiCapture software includes image pre-processing, document classification, data and text extraction and execution of validation rules. All the operations are performed automatically in the unattended mode on a server.

Image Pre-Processing 
The imported document images might need preliminary processing in order to provide the highest possible recognition quality. Image pre-processing includes such operations as:

• Correction of page orientation

• Inversion

• Conversion to black-and-white

• De-skewing

• De-speckling

   

Automatic Document Classification
FlexiCapture identifies the document type and assembles one- and multi-page documents out of the mix of pages using advanced ABBYY technologies, which allow automatic classification of documents with variable layouts of any complexity including:

• Multi-page documents

• Documents with variable number of pages

• Documents containing multi-page tables

• Documents with image or text attachments

   

The classifier can be trained automatically on a package of unsorted images requiring minimal participation from the user. The classifier is based on similarity detection, and is able to identify classes of documents and automatically find documents which belong to each class without any assistance.

Automated Data Capture and Text Extraction
ABBYY data capture software accurately extracts data and text from the fields specific for each document type using ABBYY award-winning multi-language recognition technologies. It offers:

• OCR for up to 190 languages including: Chinese, Japanese, Korean, Vietnamese, and Hebrew

• ICR (intelligent character recognition) for hand-printed text for over 110 languages

• OMR (checkmark recognition) for a wide spectrum of checkboxes

• Barcode recognition for a variety of 1D and 2D barcodes

   

ABBYY data extraction software is easy to configure and maintain. The administrator can "point-and-click" data fields for extraction on a training image set for further production level document processing. In addition, there is a unique capability of fine-tuning document descriptions in order to reach unsurpassed accuracy in locating and recognizing document fields.

Automated Validation of Data

FlexiCapture provides a set of predefined validation rules to ensure data correctness. By applying these rules data can be automatically checked against a database, checked to satisfy the format, checked if the sum is correct (for invoice capture tasks), brought to a standard (normalized) format and more. Any custom rule can be added using scripting language.

Request an onsite trial or WebEx demo