Decipher IDP 1.2: August 2021

Click this icon on the toolbar to view and download a PDF version of the release notes.

Decipher IDP requires the following components, available for download on the Blue Prism customer portal:

Blue Prism

6.6.0 or later

This release of Decipher IDP is compatible with these versions of Blue Prism.

Decipher Licensing Service

N/A

The Decipher Licensing Service.msi is not versioned for this release.

Decipher Server

1.221.03230

The required Decipher Server version for this release.

Decipher Web Client

1.221.05130

The required Decipher Web Client version for this release.

Decipher Automated Clients

1.220.12070

The required Decipher Automated Clients version for this release.

Decipher VBO Interop dll

N/A

The BluePrism.Decipher.VBO.Interop.dll is not versioned for this release.

Decipher Blue Prism business object

N/A

The BPA Object - Decipher.xml is not versioned for this release.

Decipher Server Plugin

N/A

The Decipher Server Plugin.msi is not versioned for this release.

Known issues

A list of any prominent issues with this release is maintained in the knowledge base – click here for more information.

Important notices

A valid Decipher IDP license file is required to activate the software. See Decipher IDP licensing for details.

Decipher IDP licenses issued prior to this release are not compatible with Blue Prism 7.0 and later, these licenses will need to be upgraded.

Upgrading

If you already have a version of Decipher IDP installed, please review the Upgrade notices before installing this release.

Secure development policy

As part of our secure development policy, Blue Prism products are evaluated against the following security standards:

  • OWASP Top 10 –The security vulnerabilities that pose the most risk to applications development. For more information, see the OWASP website.
  • SANS CWE Top 25 – The most common software security vulnerabilities found throughout the software development industry. For more information, see the SANS website.
  • Federal Information Processing Standard (FIPS) Publication 140-2 – U.S. government standard that defines minimum security requirements for cryptographic modules in information technology products.
  • Payment Card Industry Data Security Standard (PCI DSS).
  • Health Insurance Portability and Accountability Act (HIPAA).
  • Federal Information Security Management Act (FISMA).

Enhancements

Description of change

A new Natural Language Processing (NLP) plugin is now available for Decipher IDP. Once installed, Decipher IDP document types and batches can be configured to recognize natural language text. The NLP plugin requires a graphics card with a Nvidia Graphics Processing Unit (GPU). This release of the NLP plugin recognizes English text only. Learn more...

Users can now apply format-specific variants to Document Form Definitions (DFDs) when editing a new or existing DFD. This means that single DFDs can now be configured to recognize and capture format-specific data from batches of the same type of document, but with differing structure or content. A new Specific versions drop-down has been added to the add and edit Document Forms Definitions screens, with Create new and List specific versions options. Learn more..

The following enhancements have been made to the Document Form Definition screen accessed when editing a new or existing DFD:

  • The SQL name button has been removed and the SQL table name functionality is now accessed via a new settings button. The settings button is on the top right-hand side of the screen and is depicted with a cog icon.
  • The Add table option is now accessed via the ... (ellipsis) button.

  • The # column has been removed, this displayed the position number of each row and remained in focus when the user scrolled across the field definition options. Instead, the Title column now remains in focus while the user scrolls to view the field definitions.
  • The Document form definition name field has been removed. Users can now edit the DFD’s name by hovering their mouse over the DFD name at the top of the screen.
  • The Mappings and Preview buttons have been moved to a more prominent position below the DFD’s name.

A new screen called Capture Models, which enables users to import, export, amend, and delete Decipher IDP machine learning models has been added to Decipher IDP. The new screen can be accessed via the new Capture Models menu option, displayed on the left‑hand menu on the Admin Panel. Learn more...

The following changes have been made to the Admin/username menu:

  • A new Settings menu option allows users to access the new Settings window to change the language of the user interface and to configure their image view settings and layout preferences to be applied during data verification.

  • The Language menu option has been removed and the options are now accessed from the new Settings window, as described above.

When installing the Decipher server, users can now choose to create a separate reporting database, and if selected, to decide whether to create that database on the same server or a different server. The new reporting database allows Decipher IDP to record the performance and analytics information separately to the core data and also enables users to retain reporting data for much longer, if required.

Any existing reporting data can be copied over from the Decipher server to the new reporting database. Learn more...

When configured with a separate reporting database, Decipher IDP offers users two new report types, Accuracy and Automation Rates:

  • Accuracy reports – Report on the accuracy rates of document classification or data capture by batch. Learn more...
  • Automation Rates reports – Report on the automation rates of document classification or data capture by batch. Learn more...

Machine learning data and field mapping data are now shared across the Decipher IDP environment and are no longer user-specific.

The following changes have been made to the Add New User and Edit User windows:

  • A new field called Site enables users to enter a specific site, or area in which the user works.
  • The Copy learning from drop-down list has been removed. This is because learning data is no longer user-specific and all users share and contribute to learning data.
  • The Copy mappings from drop-down list has been removed. This is because mappings data is no longer user-specific and all users share the same mapping data.
  • The Confirm password field has been renamed Confirm new password.

The following new fields have been added to the New and Edit Document Types screens, accessed via the Admin Panel:

  • Primary recognition language – The primary language to use when processing documents using the document type.
  • Primary locale – The primary locale is used to validate dates and parse addresses during processing using this document type.
  • Secondary recognition language – An additional language to use when processing documents using this document type.
  • Secondary locale – An additional locale for this document type.

As per existing functionality, these options can also be configured by batch type. Any document type settings will override those configured by batch type. However, if these settings are not configured by document type, the relevant batch type settings are applied.

The following changes have been made to the machine learning options on the Edit Document Type screen, visible when the Machine Learning option is enabled:

  • A new Capture Mode drop-down allows users to specify whether the selected machine learning model is being used to process structured or unstructured data.

  • A new Training Size field allows users to specify the number of documents that should be used when training the selected model.

  • New Periodic training options allow periodic training to be configured to run automatically after a specified number of documents, or it can be disabled and the Train now option can be used.

  • New Train now functionality informs users when the model was last trained and the number of documents that have been processed using this document type since the model was last trained. Users can also choose to train the model now by selecting Mark for training.

    The Mark for training check box is not automatically unselected after the training has been completed and must be unselected manually.

  • The automatic option has been removed from the Train model settings.

Users can now import and export Decipher IDP training data from the Training Data screen. Previously, users could only delete existing training data for the current user from this screen.

A new Exceptions tab has been added to the Select a batch screen. The new tab lists the batches of documents that have been marked as an exception during data verification, and enables users to further process exceptioned batches. The tab only displays for users with the new Resolve Exceptions permission applied. By default, this permission is not enabled for any user or user role.

A new Validation errors dialog has been added to the batch validation process, which lists all of the validation errors that were raised on the batch. Previously, only one validation error was visible, making it difficult for users to mitigate the errors and allow the processing of the batch to continue.

To improve the accuracy of data capture, the orientation of document pages is now applied during OCR, rather than before. This enhancement improves the accuracy of the automatic page orientation from 90% to 99.9%.

To improve performance, large batches of template training data are now imported in small batches rather than one file.

The following improvements have been made to the Batches screen:

  • Filters have been added to the top of each column to enable users to search for items and change the view.
  • Batch ID and Current user columns have been added.

During data verification, users can now press Alt + X on the keyboard to temporarily hide all existing regions. This enables users to add a new region in a non-cluttered space. The regions re-display when Alt + X are not being pressed.

Users can now edit batch types directly by double‑clicking on the batch type on the Batch Types screen.

The following usability and consistency improvements have been made to the Decipher IDP user interface:

  • Help buttons have now been added to all screens.
  • During data verification the region re-size handles now adapt to the user’s screen resolution.
  • The table rows in the Select a batch window are now blue rather than yellow to use ensure consistency throughout the user interface.

  • The About button in Data Verification has been renamed to Help.

It is no longer necessary to log out and back into Decipher IDP to update the menu options when toggling Class Verify or Data Verify on and off on the Configuration screen.

By default, the size of the Formula dialog, which displays when a user modifies a Formula field in a DFD is now bigger and can also be expanded further.

To improve performance and to ensure up-to-date training data is always available to Decipher automated clients, a new ‘selective update’ mechanism for training data has been introduced for the Decipher server. Automated clients are now automatically updated with training data changes as they occur. Previously, all of the data was refreshed at set intervals, regardless of whether the data had changed. This does not apply to mapping definition changes – this will be addressed in a later release.

The performance of the Decipher server, when communicating with a remote database on systems processing high volumes of data through multiple automated clients, has been improved. Automated, non-batch process data is now cached to reduce unnecessary communication between the server and the database.

Fixes and minor improvements

Description of change Reference

In systems with multiple active users, batches of documents can now be successfully added from Blue Prism to Decipher IDP. Previously, an Advanced Message Queuing Protocol (AMQP) operation is interrupted message displayed when ten or more users were performing actions in the Decipher IDP web client at the same time, preventing batches from being added.

DEC-254

An issue during data verification where Decipher IDP repeatedly failed to recognize punctuation in the same number has been resolved. Previously, commas between some numbers, such as 10,000, were not being captured.

DEC-267

The machine learning model specified in a document type is no longer intermittently being incorrectly replaced by a different machine learning model.

DEC-268 (BG-7341)

An unexpected server error, which displayed when logging into Decipher IDP if the system attempted to load before all of the required services have successfully started, has been replaced. More specific warning messages, which indicate the cause of the issue, now display to enable the issue to be better mitigated by the user.

DEC-284

An issue has been resolved where Decipher IDP repeatedly extracted currencies incorrectly, for example 100 rather than 1,00, in documents containing more than one currency, despite repeated verification. The region drawn around currencies has been enlarged to ensure all characters are now included in the extraction.

DEC-267

When editing or creating a document type, if a user tries to apply a machine learning model that differs to the model already applied to the associated DFD through a different document type, a warning now displays. This informs users that changing the machine learning model for the current document type will change it for all other document types that are associated with the same DFD.

N/A

An access violation error, which intermittently occurred for multi-device environments during data capture, has been addressed.

DEC-265

Decipher IDP now accurately extracts light and multicolored text during optical character recognition (OCR). Previously, such text was not always recognized, which meant that the data verification process was unable to create accurate regions, or map extracted text to the document as expected.

DEC-245

An issue has been fixed where the Reports Dashboard continuously flickered after the user repeatedly changed the magnification of the screen. This issue only occurred in systems with large volumes of data.

DEC-249

When a user changes the status of a batch from Exception to Open, Decipher IDP now checks that image processing has already taken place on the documents in that batch. If image processing has not taken place, then the batch is added to the image processing queue. Previously in this scenario, such batches were not sent for image processing and if this hadn’t already taken place, no data could be extracted.

N/A

An issue has been addressed where filtering very large static validation lists during batch type configuration caused out of memory exceptions. Decipher IDP can now handle static validation lists containing up to one million items.

N/A

An issue has been addressed where the OCR accuracy of non-English text extracted from scanned PDFs was not as high as for vector PDFs. Such data is now passed through a dictionary, using the language defined in the document and batch type settings, before being processed via the OCR client.

N/A