ICR vs. OCR
ICR and OCR are both designed to recognize scanned text, but there is a fundamental difference in the algorithms that these technologies use, and the types of documents that can be processed using each. OCR is the better-known technology, having existed for many years and established itself as an essential tool in many offices. OCR, which stands for optical character recognition, uses the optical features of text characters to recognize scanned printed text. The important defining feature of OCR is that it is restricted to the recognition of printed text, and cannot handle handwritten words. ICR was developed to deal with this limitation. ICR, an abbreviation for intelligent character recognition, can adjust to handwriting variation in order to recognize handwritten text.
Combining ICR and OCR in a single program
ICR and OCR are both impressive on their own, but when combined into a single program, these tools enable the automatic recognition of almost any text that might be found in a scanned document. With the exception of cursive text, which cannot yet be recognized with sufficient reliability, and letters, numbers and symbols in a scanned document can become readable with a combination of ICR and OCR. These two technologies can be combined into document automation software, designed for the processing of scanned forms.
Documents for ICR and OCR
Depending on the contents of a scanned document, it can be processed with OCR, ICR, or some combination of the two. It is common to find documents that require nothing more than basic OCR. As long as all of the text in a document is computer-printed, basic OCR should be sufficient. Some examples of documents that can be processed with OCR alone are invoices, EOBs, and education records. However, as soon as handwritten text is introduced, ICR also becomes necessary. Many forms, such as applications, surveys, and even bank checks contain handwritten information that previously had to be processed manually. However, all of these documents can now be processed with a combination of ICR and OCR.
[ Back ]