At the Ephesoft Innovate conference in October, 2016, the Ephesoft team highlighted the new features that would be part of the Ephesoft 4.1 release. Ephesoft 4.1 is now available, and this article will highlight some of the main new features in this version.
The first major change is that the Ephesoft product has been re-branded from “Ephesoft Enterprise” to “Ephesoft Transact”. Moving forward, there will be “Ephesoft Transact” and “Ephesoft Universe” products.
At Ephesoft Innovate, the Ephesoft team described their focus on the following areas for the Ephesoft Transact 4.1 release:
- Accuracy
- Productivity
- Security
- Connectivity
In the rest of this article, we will highlight several new features that fit into the focus areas listed above.
Accuracy
Several new features are now available to improve the accuracy of both document classification and extraction.
Paragraph Extraction
Ephesoft now has a way to pull values locked inside of multi-line text fields. With the new release, it is possible to look for a key target phrase to start looking within a paragraph to find specified values within a block of text. Using the key target value, Ephesoft will know when the targeted paragraph data has started, and then one or more values can be pulled from this block of text.
Along with pulling values out of a block of text, it is now possible to have wrapped data extraction, or multi line extraction across a line break.
Multidimensional Extraction
Ephesoft has long provided key/value extraction where the “key” and the “value” were set to target locations on a given page. This greatly improved extraction accuracy, as it was possible to specify a region on the page where a key would appear, as well as the location of a potential target value.
Ephesoft 4.1 takes that one step further with cross section extraction. It is now possible to provide an horizontal key and a vertical key that point to a target value at the intersection of those two fields. For example, if a specific field in a table needed to be captured, a row element could be used as the horizontal key and the column header could be used as the vertical key. This can improve the accuracy of extraction of a target field.
Multiple Barcode Engines for Extraction
For a given page with barcode data to extract, it is now possible to specify different barcode engines to use for different barcode types on the ame page. If a page has multiple barcode types, that means it is possible to specify which engine should be used for each extraction.
Machine Learning for Improving Document Classification
For Classification (determining the document type of a given document), Ephesoft has provided a way for Batch Class administrators to “train” Ephesoft as to the first, middle, and last pages of a particular document type. Once trained, additional samples can be added to improve classification accuracy as needed.
With Ephesoft 4.1, this has been taken one step further: Users can be allowed to add samples to better train classification from the “Review” pages. A simple pop-up window allows a user to drop example first, middle, and last pages as classification examples. This allows for ongoing improvement in document classification.
Classification by Multiple Methods
A new plug-in has been added in Ephesoft 4.1 that will provide the ability to use more than one method for determining document classification. If more than one method is used, Epheosft will determine the one that has provided the highest classification confidence, and use that to determine how a given document should be classified as a document type.
Productivity
Along with providing improvements to the accuracy of classification and extraction, Ephesoft Transact 4.1 provides several new features to improve user productivity.
Automatic Regular Expression Suggestion and Creation
Ephesoft relies heavily on the power of regular expressions for extracting and validating data extracted from documents. For users new to regular expressions, it can be daunting to come up with the correct syntax for more complicated expressions. Ephesoft now provides suggestions for users in a pop-up window based on selected text. This allows users to select example field values, and then immediately see a suggested regular expression for that extraction.
Formatted Field Extraction
After setting up the extraction rules for an index field in Ephesoft, it is now possible to configure automated formatting rules for the extracted data. This feature is easy to set up for a field via a screen in the Batch Class Administration pages. Tools are provided to easily setup common data transformations:
- Add a prefix and/or suffix to the extracted data
- Perform text replacements on the extracted text
- Change the case of the extracted field (to either all uppercase or all lowercase)
- Fetch a named group by a regular expression
- Date formatting and conversions
The configuration screen also provides a tool for testing sample data during configuration, so it is easy for a user to validate their formatting rules after they have been created.
Automated Line Item Matching
Ephesoft can perform table extraction, and perform validation against the data that is extracted. However, users often need to compare the extracted line items against another table. The classic example for this use case is comparing the line items of an invoice to the original purchase order.
Ephesoft now provides tools to automatically match up extracted rows from an invoice table against a database table containing rows from another datasource (such as a purchase order). Lines that do not match up are highlighted for quick review by the user.
Fuzzy Database Enhancements
Fuzzy Database configuration has been moved from being configured as a module plug-in to being configured at the document level. It is also now possible to perform multiple searches against the database tables.
Connectivity
There are several new supported platforms with Ephesoft Transact 4.1. The following list shows some examples of newly supported platforms:
- When using Linux, a Fixed Form Designer tool is now available. This was previously only available on the Windows installation.
- Support for Oracle databases has now been added
- An export plug-in is now available for SharePoint 365
Security
Some important security updates have been made in this new release as well:
- For integrating with a scanner attached to a user’s computer, an HTML5 scanner service is now utilized (The NPAPI and ActiveX plugins have been removed)
- PIV/CAC based authentication is now supported
- Single Sign On (SSO) with SAML is now supported
Conclusion
The Ephesoft Transact 4.1 is now available, and it provides a variety of new valuable features to improve the speed and accuracy of processing documents. This article provided an overview, but Blue Fish will also be working to provide deeper dives on each of these new features.
