Black Ice Software is proud to announce the release of the TIFF Viewer 12.10 now using Tesseract 4.0’s recently redesigned OCR engine.

The Tiff Viewer OCR engine now allows for three training files (data sets): Fast, Best, and Legacy. Each Training file is over 30MB. Due to the large file size, the Tiff Viewer installers will include only the English version of the Fast training file.

As the names of training files imply, utilizing the Best OCR training file will provide the highest recognition percentage, the Fast training file will provide the fastest OCR recognition speed, and the Legacy training file performance is between Fast and Best. In addition the Legacy training file provides backward compatibility for Tesseract OCR engine 3.6.

Based on Black Ice Software QA testing, using English language documents, there was found to be a significant character recognition improvement with Tesseract 4.0 approaching the recognition performance of MODI. The OCR recognition speed was found to be 40% faster, depending on hardware configuration. Testing found the Fast training file produced the best test results in both speed and accuracy, while, the Best training file was still significantly slower. The Fast training file recognition rate, with our test documents, reached 99.80% accuracy compared with the old OCR engine scoring 94.70% accuracy.

The Black Ice Software’s OCR Language Pack installer will include all three training files (Fast, Best, Legacy) as well as every supported OCR language. Users can experiment with the different training files to confirm which file provides the best result for the document type they are utilizing most frequently. The Best and Legacy training files can be downloaded from the Black Ice Software Website by clicking on the download link for the Language pack installer here.

Technical detail summary from Tesseract’s Web site:

• The Fast training file, fine tuning/incremental training will NOT be possible.
• The Best training file is for people willing to trade a lot of speed for slightly better accuracy. It is also better for certain retraining scenarios for advanced users.
• The Legacy training files option provides backward compatible and uses a smaller neuron network.

Free Upgrades and Priority Support

Do not forget to subscribe to future releases by purchasing maintenance. With an active maintenance subscription, you will receive all new releases for free! Priority email technical support is also included in your subscription. For more information, call us at +1 (561)-757-4107 Extension 3, or sales@blackice.com.

As a special offer, if your maintenance subscription has expired or you did not include the maintenance subscription in your original purchase, simply refer to this newsletter to renew or add maintenance retroactively!

For license information, please visit our Customer License Management Portal.