With the latest releases of Tungsten Automation (formerly Kofax) TotalAgility® and Capture, Tungsten is including Tungsten Transformation 7.0. One of the key changes with that release is that Tungsten has deprecated three of the recognition engines—the software that converts images of text to processable text data—you may be familiar with and may be using.
As of this release, Transformation no longer includes the FineReader, RecoStar, and the separately licensed A2iA engines. Tungsten replaced them with their own OmniPage engine and a separately licensed Parascript FormXtra engine choice for situations previously using the A2IA engine. Operation of the separately licensed, cloud-based Clarity engine stays unchanged.
OmniPage becomes the new standard engine for both zonal and page OCR profiles in Transformation. Other than missing support for a rare RLE (run length encoded) Tiff format, OmniPage should be a direct replacement for either FineReader or RecoStar.
As noted above, the Clarity engine is unaffected by this change. In case you have yet to experience it, Clarity, in its Text Mode, performs full-page recognition and is specialized to recognize text at any location on an image. For example, photographs of people or places, street signs, ID cards, or business cards. Because these images can be distorted, have perspective issues, or contain glare from reflected light, Clarity overcomes many of these issues and provides successful recognition results. Clarity also recognizes classic documents and forms when configured to use Document Mode.
The Parascript FormXtra engine recognizes handwritten text within a zone. As part of the zone definition, one can specify the expected type of handwriting like cursive, handprint, or boxed handprint. Added settings in certain field types help further refine extraction. A set of image cleanup settings also helps improve results.
Make certain you add time in your next TotalAgility or Capture upgrade effort for transitioning to the new recognition engine options. Our experiences show very few difficulties with the change. Testing your transformation models after the switch, of course, is key to a successful upgrade.
If you're interested in staying up-to-date on your software, subscribe to our Insights to get the scoop on the latest releases.