Integrate with Tungsten Capture

Tungsten Transformation processes unstructured and structured documents by classifying them and extracting data from them.

The classification process results in a category (class) to which the document is assigned. Classification can be hierarchical, with every node in the hierarchy being a possible result.

The extraction process results in the locations and values of items on the document. These values are stored in fields and each extracted value has a calculated confidence that is an estimate of the accuracy of the value.

All classification and extraction definitions and patterns that are created in Project Builder are stored in project files (called the project). The project is used during runtime to process documents in accordance with the specified settings and to provide the results to Tungsten Capture.

Since Tungsten Transformation’ categories and fields are created and managed outside of the Tungsten Capture Administration module, they are not immediately available for processing with a Tungsten Capture batch class.

To process documents in Tungsten Capture with Tungsten Transformation, Tungsten Transformation - Server must be added to the list of queues for a batch class and the classes and fields from the Tungsten Transformation project must be synchronized with the Tungsten Capture document classes, form types, and index fields. Synchronization is performed with Synchronization Tool, which is available for any batch class that includes Server. Extended Synchronization Settings can be defined for a batch class, for example to add a project description, or to allow batch editing. Note that you do not need to run synchronization if you change those extended validation settings; you only have to publish the batch again.

When the synchronized batch class is published, the classification and extraction project data is processed in the same way as all other batch class data to provide a stable set of settings for existing batches.

If wanted, statistical data can be gathered to gain visibility of the production workflow. You can either use Tungsten Reporting or the Transformation Statistics export connector to retrieve and store information such as the time needed to classify or extract a document, the time to process a batch for Validation or the recognition field accuracy. Tungsten Reporting is a separate product that was developed as client/server application that can retrieve data from various different Tungsten products. Whereas Transformation Statistics export connector is a Tungsten Transformation functionality that is installed as add-on to Tungsten Capture. It is configured on document class level and stores statistical data using a Microsoft SQL or Microsoft Access database. Reports for the gathered data can be viewed by the Tungsten Transformation - Statistics Viewer. For more information about Tungsten Reporting, see Tungsten Transformation for Tungsten Reporting Getting Started Guide.