Our ICR recognition engine is size independent and is able to read every handwritten style: constrained and unconstrained. There are versions suitable for different character sets:

American Digits
European Digits
Uppercase Alphabet
Lowercase Alphabet
Mixed Alphabet

You can recognize single characters, words, lines and full paragraphs using the several functions available.

Accuracy: ICR recognition engine's accuracy is 95-100% in case of good quality typed and scanned images. Even with low or medium quality images, the error rate is usually very low.

Technology: our engine is based on an artificial neural network and on fuzzy-logic technology.

Speed: speed is linked to cpu speed. Using Pentium IV processors, time performance are usually up to 150 CPS.

Input: monochrome bitmaps at resolution of 200 or more DPI are accepted in input.

Output: output is made up by recognized characters as well as by confidence levels. Placeholders are used for unrecognized or very suspicious characters.

Platform: the engine is packaged in DLL format. For each content type (American Digit, European Digits, Uppercase Alpha, Lowercase Alpha, Mixed Alpha, etc...) is required a separate .REC file. The full engine size is less than 5 MB.

Only for V.I.P