Intelligent character recognition

Electronic Character Recognition ( ICR ) is an extension of the optical character recognition (OCR). She is - roughly speaking - a handwriting recognition system that can learn by computer different forms of the medium during processing. ICR software is often based on adaptive programs that rely on neural networks. New learned manuscripts find direct equivalents in its database. These specially aligned, electronic text recognition enlarged the scope of existing scanning devices to the processing of handwritten documents. OCR covered so far only from the area of ​​processing of printed documents in a satisfactory manner. Since the process of handwriting recognition designed much more demanding, but so far the achievable accuracy here was inadequate. In well- structured documents are possible with ICR hit rates of over 97 %, plus the software requires multiple passes in alternating methodology. Each of them is weighted differently in the evaluation. Even separate methods for the numbers and the letters of recognition are used. An important step for the ICR has been the development of so-called automated forms processing in 1993. This describes a three-step process for the inclusion of a character or an image by the ICR software. In the first step an image of the document is created. This recording is processed in the second step of the ICR software and finally automatically evaluated in the last step.

Electronic text recognition increases the efficiency of the optical character recognition significantly. ICR software found in the business world, for example in the processing of handwritten filled out forms application. ICR solutions are available from the following manufacturers:

FlexiCapture Engine

ABBYY FineReader Engine

OpenRTK