Vantage application #: 0.46.0
Advanced Designer #: 0.53.0
Platform #: 0.30.2
Technology Core #: 2.31.2
Vantage Solution #: 0.22.3-onprem
Support for training of repeatable barcode fields in Document Skill
Resizable Script Editor Window
Command Continue table from this row is applied to 10 pages
Updates to existing built-in skills
- Improved the quality of extraction of companies' names, addresses, and PO numbers.
- Additional languages are now enabled for the Invoice EU skill.
Improved the quality of Line Items extraction.
- Improved the quality of all fields for supported countries including Sweden, Norway, and Finland.
- Additional languages are now enabled for the Receipt skill.
Trained on new documents, quality improved.
New field: Driver License Restriction Code.
Form W-9, Request for Taxpayer Identification Number and Certification
Can now process handwritten forms.
Form 1098, Mortgage Interest Statement
Improved the quality of the skill.
IRS Tax Form 1040, U.S. Individual Income Tax Return
New skill (IRS Tax Form 1040, U.S. Individual Income Tax Return, 2022) created to process ISR Tax Forms 1040 for the latest year. The IRS Tax Form 1040 process skill (with a corresponding classifier) has been created to process Forms 1040 for different years simultaneously.
New built-in skills
- Form 1099 Composite
- Brokerage Statement
- Form 1099-B, Proceeds from Broker and Barter Exchange Transactions
- Form 8949, Sales and other Dispositions of Capital Assets
- Form 1095-C, Employer-Provided Health Insurance Offer and Coverage
- Form 1095-B, Health Coverage
- Insurance Health Card US
- Business Card CJK
- Export extraction results into CSV
- Export document images into TIFF, JPG, and non-searchable PDF files
- ICR for Spanish
- Further improvements to ICR for French
Improvements to OCR Technologies
- Further improvements to the quality of Japanese OCR
- OCR quality on Documents with Watermarks
- OCR for some European documents
- Increased OCR speed for uncropped photos
- Technology core 1.0 end of support
- Manually crop and split documents at manual review
- No "Send to Stage..." button in the Manual Review interface when integrated using iframe
- Special interfaces for capturing ID cards
- Disabled uploading images from gallery
SAML 2.0 protocol support for authentication
Using the same external identity provider in several Vantage tenants
Single sign on experience with external identity providers
ABBYY Vantage Advanced Designer
- Deep learning statistics
- Accessing data catalogs from extraction rules
- Working with tables and repeating groups
- Searching for vertical text only in extraction rules
- Using cross-validation to estimate accuracy of NLP Segmentation activity
NLP Language support
- In the "Select Skill" dialog, skills were not sorted alphabetically.
- Documents with uncertain symbols were not routed to Manual Review when the condition "With rule errors and uncertain fields or unknown" was selected.
- The extraction quality of Handwritten fields was insufficient.
- The "Task reject" action was not described properly in the documentation.
- No notifications were displayed when a catalog update was completed.
- An error occurred when trying to match a modified Identity Document skill.
- The number of documents to be deleted from a skill's document set wasn't shown correctly in Japanese UI.
- The "Custom" Data Catalog type was not described correctly in the documentation.
- Text was missing on Arabic Identity Documents.
- The order of pages sent to Vantage from a Scan Station wasn't preserved.
- It was not possible to publish a skill after it was trained successfully if any errors occurred during the training process.
- An error occurred when training the InvoiceCA fields.
- The quality of recognition results decreased after upgrading to v2.3.
- When opening Manual review and then navigating to another tab, the document image disappeared.
- An error occurred when extracting after training.
- The quality of the vendor extraction results worsened after upgrading to version 2.3.
- The "Name/Full name" field was recognized incorrectly on Hong Kong ID cards.
- The program stopped responding when loading a data form in the Document Skill Editor for some skills.
- When clicking a link on the 6th or later documents in the Document set, the focus in the batch editor was set to the 1st document.
- The failure of a publish operation caused the workspace to become inconsistent.
- A field populated with a script rule was indicated to be located incorrectly in the Results tab of Document Skill Editor.
- The PDF text layer was recognized incorrectly on some documents.
- The Skill Monitor erroneously displayed a "Started" status for transactions that should have had the "Completed" status.
- The scrolling area was too narrow.
- Upgrading a copy of an Invoice skill to the latest version caused the program to become unresponsive.
- An "InvalidOperationException: Block with ID ... already exists on page..." error occurred when uploading a PDF document to the Process skill.
- The documentation incorrectly stated that a regular expression defined for a field was used for field value validation.
- An error occurred after authenticating in Vantage Scanning Station.
- An error occurred while processing some PDF files.
- Document processing failed if a checkmark group had checkmarks on different pages.
- PDF images were erroneously flipped horizontally by the Identity Document Skill.
- It took 10 seconds for Fast Learning to process a document.
- For Process skills containing more than one Custom Activity, export to PDF and OCR repeated several times.
- Line items were duplicated.
- An error occurred while setting up using Azure Active Directory (AD) as the external identity provider.
- In Vantage Skill Monitor, page recognition failed, causing the workflow to not be completed.
- An error occurred when processing PDF files using the OCR Skill.
- A Custom activity failed if multiple regions were selected for one field.
- The documentation erroneously stated that there were extra languages that could be enabled for built-in Invoice skill.
- Online learning didn't work for some skills.
- Output SFTP settings weren't editable in a copy of a locked Process Skill.
- Manual Review image scaling wasn't saved across multiple transactions.
- The Document skill training took too long.
- Extraction results of the Process skill differed from the results of the Document skill.
- Documents with uncertain symbols were not routed to Manual review when the condition "With rule errors and uncertain fields or unknown" was selected.
- Document images weren't displayed when a task was automatically opened in Manual Review.
- Tenants could access documents of other tenants.
- The ImageWidth and ImageHeight properties of the Page object returned 0 when they were used in the Custom activity script.
- An error occurred while capturing line items.
- The ID Reading skill produced a date in US format.
- The output XML file of an OCR skill didn't contain the "rotation" attribute of the "Page" element.
- Data forms were automatically scrolled down after a task was opened at the Manual Review stage.
- Document processing failed on classification or extraction.
- An error occurred when changing the document type in Manual Review.
- Skill versions were ordered incorrectly in a Processing Skill.
- The tooltip for the Context.Transaction object didn't contain the Parameters property.
- An error occurred when importing TXT files into Vantage.
- A "Max Activation Count..." error occurred when processing documents in Vantage.
- The tenant administrator's name was displayed in the Security Log for OML events.
- A blank page was shown instead of the document when working with line items.
- Documents were not added to the Document set if there was a Custom Activity after Manual review.
- An error occurred when trying to search a catalog.
- An error occurred when training a Classification skill.
- The source file name wasn't displayed for imported skills in the Skill Monitor.
- An incorrect message was displayed in the Security log online learning if the model was not applied.
- An error occurred when trying to review the training results of a document skill if there were some fields without regions.
- Fields’ OCR settings were not applied if the skill had no Fast Learning model, for example, if it was created in Advanced Designer and did not contain a Fast Learning Activity.
- Adding a Fast Learning activity to a Document Splitter skill prevented the skill from working.
- Uploading a CSV file to a Data catalog failed if there was a space character in the catalog name.
- Advanced Designer. Neither the Script nor the IF activity editor opened after an empty group has been created in the Manage Fields dialog.
- Advanced Designer. When restoring a skill with duplicated documents that was saved to a local folder, the restored skill would contain just one copy of the duplicated documents.
- Advanced Designer. A group created on the Data Form pane turned into a table when the Allow Multiple Items parameter was enabled in the Manage Fields dialog.
- In Advanced Designer, adding a long comment (more than 256 characters) to a document in an Extraction Rules activity caused an internal error.
- Field type limitations for the Address Parsing activity were not specified in Advanced Designer documentation.
- In Advanced Designer, fields that were extracted successfully when matching an Extraction Rules activity weren't extracted when testing the skill.
- There was no warning that a copy of the Identity Documents skill can't be edited in Advanced Designer. The user could try to open a copy for editing, but loading froze.
- Advanced Designer could be blocked by a customer's IT infrastructure because it used an outdated version of Microsoft Edge WebView2 runtime.
- In Advanced Designer, the X icon for deleting a field region on the Fields tab was located below the area where the user should actually click to delete a region.
- In Advanced Designer, the Deep Learning for NLP activity didn't support text fields with data types Date and Number. Additionally, field type limitations for this activity were not specified in Advanced Designer documentation.
- When adding a new item for a field with multiple items, the region corresponding to the previous field item remained selected on the document image, meaning the user couldn't specify a region for the newly added field item.
- In Advanced Designer, it wasn't possible to review a recognition issue by clicking a value in the results table on the Results tab of a skill or any of its activities.
- When creating a skill from a local folder, only unique documents were uploaded to Advanced Designer. If the document set contained duplicate documents, only one instance for each document was uploaded.
- There was a misprint in a script sample for the Splitter Script activity in Advanced Designer documentation.
- In Advanced Designer, when data extraction was configured using an Extraction Rules activity and a lot of similar hypotheses were formulated for search elements, data was successfully extracted when matching the activity but wasn't extracted when testing the activity.
- In Advanced Designer, fields that got regions from optional Region search elements were extracted successfully when matching the Extraction Rules activity but weren't extracted when testing the skill.
- In Advanced Designer, an Extraction Rules activity couldn't find Separator search elements when the penalty length was set to the default value.
- A skill that was created in Advanced Designer 2.2.1 or earlier versions but wasn't published before updating to Advanced Designer 2.3 or later, couldn't be opened for editing after updating Advanced Designer.
- A user could try to save a skill locked by another user to a local folder, which caused an internal program error in Advanced Designer.
- Unless an option was selected from an opened context menu for a field with multiple items, the context menu could not be closed and remained permanently on the screen.