Text Recognition

Our book scanning process is executed in three parts. The first is capturing the images of the pages. The second is pre-processing the images for optimal text recognition. The final step is the OCR and final file conversion.

Test Results

To illustrate the the accuracy of Blue Leaf's commercial-grade OCR software, we scanned a low quality text image and processed it using both Adobe Acrobat Pro's (CS3) OCR engine as well as Blue Leaf's engine. The test result are provided below.

Character Errors*
Adobe Acrobat Pro: 9
Blue Leaf: 1

The results speak for themselves. Blue Leaf uses the same OCR engine as many of the major Fortune 500 and IT companies in the United States.

*Please be aware that OCR technology is not perfect. The results of this test are for comparison only, under controlled conditions. Actual OCR accuracy may vary from book to book, depending on print quality, presence of rare characters such a Latin diphthongs, complexity of font styles (the more ornate the font, such as script, the greater likelihood of read-errors) and some italicized characters.

If the customer is planning to publish the resultant text files, then it is expected that he will need to proofread the text for OCR errors, prior to submission. For more details, please read our FAQ.

OCR Comparison

Blue Leaf OCR

Acrobat Pro OCR

ERRORS FROM LOW-RES TEST SAMPLE
BlueLeaf	Acrobat Pro