PrimeOCR as an Capture/InputAccel Module |
|
Both PrimeOCR and PrimeVerify Software are available as Capture/InputAccel - Certified Modules.
Prime Recognition has added Capture/InputAccel compatibility to both its PrimeOCR "Voting" OCR software and PrimeVerify software. Both products can accept and return files residing in the Capture/InputAccel Management Subsystem. Capture/InputAccel customers now have access to production oriented OCR software that generates 65% to 80% fewer errors than conventional OCR and a true "heads-up" style error correction solution for verifying OCR results.
PrimeOCR for Capture/InputAccel
PrimeOCR for Capture/InputAccel allows images captured through Capture/InputAccel to be routed to a PrimeOCR NT/2000/XP PC for high accuracy machine print recognition. Designed as a Visual Basic Plug-in Component, PrimeOCR for Capture/InputAccel can easily be customized to fit users specific OCR processing needs. PrimeOCR is a Windows NT/2000/XP based software OCR engine and Application Programming Interface (API) toolkit. Using a combination of proprietary "Voting" algorithms and technology licensed from the best retail OCR vendors, PrimeOCR is able to reduce OCR errors by 65-80% over conventional OCR products.
PrimeOCR - The Benefits of Higher OCR Accuracy
PrimeOCR reduces the cost of labor and workstations associated with manual error correction by 65-80%. Since manual error correction can account for over 50% of the annual imaging system cost in OCR-intensive applications, installations that use PrimeOCR can achieve annual cost savings of 40%.
And PrimeOCR can save valuable time even if you dont perform any OCR error correction and instead use a "Fuzzy" search engine to make up for OCR mistakes. The more accurate your OCR, the more likely the data you are looking for will appear at the top of a fuzzy searches "hit" list. This saves time by allowing you to find your data faster. This is especially true for databases with over 1,000 documents, where the number of false hits can be so large that reviewing all search results is too time consuming.
PrimeOCR Production OCR Features:
-
PDF Support in PrimeOCR - PrimeOCR lets users convert scanned images into PDF "Normal", "Image + Text", or "Image Only" formatted files. PrimeOCR's PDF output retains the original image and text layout while minimizing PDF file size with full compression. In addition, PDF output from PrimeOCR can be sent to Capture/InputAccels PDF Export Module. And, because of PrimeOCR, the PDF output contains the most accurate OCR results available to the production imaging marketplace including the ability to process color or grayscale images..
-
Cleaner Data - Only 40-60% of errors generated by conventional OCR products are "flagged" for correction. Since manual editing typically corrects only flagged errors, up to 60% of the errors produced do not get corrected. By reducing the total number of OCR errors, and generating more accurate suspicious character flags, PrimeOCR can reduce the total number of errors that make it into your database or final application by 75%.
-
High Fault Tolerance - A very poor image may cause an OCR engine to "crash". Since the production OCR system often runs unattended, PrimeOCR includes an "Automatic Engine Recovery" feature that reinitializes any failed engine. The offending image is saved to an error directory and processing continues with the next image.
PrimeVerify for Capture/InputAccel
The PrimeVerify application allows users to simultaneously edit PrimeOCR results and view the original image for more effective error correction.
PrimeVerify for Capture/InputAccel is designed to access PrimeOCR results that are residing in the Capture/InputAccel subsystem. After performing OCR cleanup, PrimeVerify users can then save corrected results back into Capture/InputAccel for export or further processing.
PrimeVerify Features
One of the best features in PrimeVerify is the ability to see
both recognized text and the original image together for easy "heads-up" style
editing. To reduce the amount of searching required to verify text, whenever text is
selected, a corresponding box is drawn around the area on the image where the original
text can be found. For easy comparison, the image is moved to center the original text
above the recognized text. In addition, PrimeVerify supports editing of single or
multi-page Tiff files and includes a number of other features that promote effective error
correction:
-
Low Confidence Character Highlighting - Suspected errors, marked by PrimeOCR, can be displayed in color for easier correction.
-
Easy Text and Image Navigation - Move through text using key-strokes or mouse clicks to locate the characters you wish to edit. You can quickly move between highlighted text using the TAB key.
-
Fast Text Editing - Select and replace characters, words, or whole lines using keyboard commands. To speed error correction, automated text search and edit is provided through Find and Replace commands.
-
View Image Statistics - To determine how well an image was recognized, PrimeVerify provides an Image Statistics table that shows the number of characters marked in up to 9 Confidence Levels.
-
Automated File I/O - PrimeVerify contains an "auto-process" feature that can be configured to perform a variety of File I/O tasks. For example, upon reaching the end of the current file, PrimeVerify can automatically save your results and open the next file.
-
Insertion Tables - These tables are used to automate typing over incorrect words. Users build tables of words often found to be incorrect. Whenever you come across an incorrect word, select it, then double-click on the correct word in the insertion table to do an automatic replacement.