Intelligent character recognition (ICR) is an extension of optical character recognition (OCR). While OCR identifies hand-printed characters, ICR lets computers recognize different font styles to improve their text recognition accuracy.
Both technologies (OCR and ICR), in the simplest terms, allow digital devices to read text.
Other interesting terms…
Read More about “Intelligent Character Recognition”
ICR is powered by self-learning systems known as “neural networks.” In ICR’s case, the neural network automatically updates the database attached to a character recognition system with newly found handwriting patterns. The constant flow of inputs for the computer to learn improves its ability to recognize handwritten text more accurately over time.
Optical and Intelligent Character Recognition, What’s the Difference?
Despite the similarities between OCR and ICR, which isn’t surprising since ICR is a subset of OCR, they have stark differences as shown below.
| OCR | ICR |
| Has many facets, ICR is only one | A subset of OCR |
| Primarily aims to recognize characters in printed documents | Primarily aims to recognize handwriting |
| Can’t read handwriting and complicated fonts very well | Can recognize complex writing fonts and styles |
| Can’t understand details and content | Can understand content |
Judging from the details in the table above, ICR is more advanced than OCR.
How Intelligent Character Recognition Works
ICR requires using at least two components—a scanner and a computer with an ICR software installed.
- The process starts by uploading the documents to an ICR scanner.
- The software reads and interprets the content.
- The data gets extracted from the documents.
- The data is proofread and checked for spelling errors.
- The application flags the interpretation for errors, which the user must correct, if necessary.
- When the content has been dubbed accurate, it is saved into an attached database, which other users can access.
Intelligent Character Recognition Applications
ICR programs have various uses, some of which are discussed in more detail below.
Financial Data Processing
The first ICR software, developed by Joseph Corcoran in 1993, was meant to automate form processing. An example would be payment processing in companies. In this case, forms (e.g., invoices) are scanned then processed to extract information. The data is then stored in a database, such as a company’s financial records. The organization’s finance department can then begin the payment process.
Data Archiving
It’s common knowledge that most companies are required to keep personal records (e.g., employee, finance, and customer data) to comply with auditing standards (e.g., International Organization for Standardization [ISO], etc.) or Internal Revenue Service (IRS) requirements. But keeping seven years’ worth of physical documents to adhere to IRS requirements, for example, isn’t easy. And if an organization needs to retrieve information from years ago, it would be harder to do so by going through boxes of records as, say, compared to opening an archive in a computer and typing Ctrl + F.
An ICR software can lessen the time and effort required to digitize printed documents for storage, not to mention would take up much less physical space that many offices may lack.
Handwriting Verification
Many legal investigations accept handwritten documents (e.g., wills, contracts, etc.) as evidence. But forged contracts and even signatures can be very challenging to ascertain even for the best forgery experts. And how many of us have watched movies where the forgers get away with their deeds? Or where the bad guys paid the forgery experts to sway the courts’ decisions?
In such cases, using ICR software to ascertain someone’s handwriting and signature veracity may be a better course of action. For one, the application can’t be bribed to give false testimony.
—
Today’s ICR software have yet to become 100% accurate, but over time and given as many inputs as possible to learn from, they may soon be.
