Intelligent Character Recognition (ICR) is an advanced form of Optical Character Recognition (OCR) that is used to recognize handwritten text and convert it into computer-readable digital text.
ICR uses algorithms and the latest AI to recognize a variety of different handwriting styles or fonts and improve character (text) recognition and accuracy on paper-based documents.
Once these documents are scanned, ICR is performed to recognize text, and the extracted data is stored digitally in a database or ECM system. The data can leveraged in business workflows, integrated into reports, and can be easily found through searching the ECM system.
OCR technology is significantly improved by ICR's ability to process different handwriting styles, facilitating data extraction from both structured and unstructured text documents.
With each encountered handwriting style, ICR uses artificial neural networks to improve its accuracy by incorporating any fresh, new data to expand and upgrade its recognition database.
Many of our clients depend on accurate documentation for managing consumer records, so capturing data with as close to 100% accuracy as possible is of utmost importance. This is where ICR plays a vital role. It is a simple but robust capture tool for reducing errors while also saving human resources like time and effort.
Table of Contents:
Discover how transformative AI and intelligent character recognition technology are in real-world business applications.
There are many benefits of using intelligent character recognition. Among those are:
Traditional OCR uses a “matrix matching” algorithm to identify characters using pattern recognition. The character on the document’s image may look like this:
It is compared to a stored example that looks like this:
By comparing a matrix of pixels between the character on the image and the stored example, the software determines the character is a “G”.
This seems like a good approach – but beware of the pitfalls. Because it is comparing text to stored examples pixel by pixel, the text must be very similar.
Even if there are hundreds of examples stored for a single character, problems often arise when matching text on poor quality images or using uncommon fonts.
We have talked about the two technologies, from a technical perspective. But what are actual reasons why businesses use OCR vs ICR, and vice versa?
Through training, ICR can be adjusted to recognize new handwriting styles and document changes. It's a more detailed and involved technology compared to standard OCR, as it can even flag potential inconsistencies or mathematical errors, requiring human review for verification.
ICR is a sub-set of OCR technology that can use the latest AI to recognize complex handwritten fonts and styles. By using modern AI tools like Microsoft Azure, ICR can adapt to new handwriting styles and convert handwritten documents into editable and searchable files.
Essentially, ICR systems scan documents and interpret all handwritten text and fonts, compared to handwriting databases. The software may ask the user to verify results.
To get more in-depth, here are the exact steps in the ICR process:
The ICR process begins when documents are loaded into an OCR / ICR scanner, camera or other imaging device. Any text (handwritten or printed) is transformed to paper to image form, either in JPG, PDF, PNG, or other format.
Just before ICR can begin, the document images have to be cleaned up. Hopefully, all noise or distortion is removed, lighting is optimized, and other non-text artifacts (like hole punches, staples, or lines) are removed.
The more that a document image can be improved, and the more that non-text artifacts can be removed, ICR data recognition accuracy improves greatly.
In this phase, ICR divides a document image into individual characters or text strings. This makes the next phase, feature extraction, easier to execute.
Intelligent character recognition engines work by combining both traditional and feature-based OCR techniques. The results of both algorithms are combined to produce the best matching result. Each character is given a “confidence score,” which corresponds to how closely the character pixels or features match or a combination of the two.
Even with this blended approach the typical text to OCR villains are on the attack: poor document quality, multiple font types, and different font sizes.
What is this character? Is it an "O", "0", "C" or “G”? Is it even a character or letter at all? Very hard to tell, especially for a machine.
So intelligent character recognition must make a decision and it may not make sense within the context of the word or sentence. If a human can’t read the character, then OCR will certainly have trouble.
From this point, the component features of each characters are identified, rather than by comparing pixels to known examples.
So instead of using pixels to recognize this character...
Feature matching is used. It's often easier for software to execute recognition through looking at features (like a character's size, curve, and shape) that make up a character instead of random pixels. The result is that the margin of error is less. These orange lines are a stored glyph for the letter 'G'.
So ICR software compares the shapes in it's stored glyphs to the character that was written or typed.
Features include lines, line intersections, and closed loops. ICR combines this feature analysis with traditional pixel-based processing to achieve high accuracy character recognition.
For example, an “O” is a closed loop, but a “C” is an open loop. These features are compared to vector-like representations of a character, rather than pixel-based representations. Because intelligent character recognition looks at features instead of pixels, it works well on multiple fonts and with handwritten characters.
Now you know why intelligent character recognition is an improvement over standard OCR.
Most intelligent character recognition software also use a machine learning algorithms like neural networks. These are like a recognition database that classify and store new handwriting patterns, character features, and styles.
In this step, ICR systems also use context analysis to examine words or sentences. This helps the software to compare the text to a dictionary and thus improve handwriting recognition accuracy.
Without additional context, character recognition errors make sense. Even if the character isn’t discernable, a human knows “ballboy” is an indie band from Scotland and “bollboy” is just gibberish:
The most common post-processing done by OCR engines is basic spell correction. Often, errors from poor recognition result in small spelling mistakes. All commercial OCR engines compare results with a lexicon of common words and attempt to make logical replacements.
Finally, data gets extracted from documents and structured based on rules and key-value pairs. The data is then mathematically validated (where applicable) and intelligently checked for errors.
When errors or anomalies are found, the ICR system, like Grooper, flags them and sends them to a queue for a human operator to review. The documents are processed and extracted document data is automatically entered into downstream business systems (such as ECM or ERP), databases, or accounting systems.
ICR software recognizes on the character level, whereas IWR looks at full words and phrases. It also has the capability of capturing unstructured data and is a different evolution of hand-written ICR.
This is not to say that intelligent word recognition is going to or should replace traditional OCR or ICR systems, as IWR is optimized for processing real-world documents that have a free form and hard-to-recognize data fields and as such are not suitable for ICR. (See example above).
So the best application of IWR is on documents where OCR or ICR would have a very tough time (cursive handwriting); but also with enough instances where the other option (manual entry) would involve substantial manual work.
There are three easy ways to solve this:
Intelligent character recognition was created for the purpose of real-world data capture off physical documents and intelligently converting it into usable electronic forms. ICR is being used every day by these industries:
One great example is shown in this case study on medical document data extraction. In this instance, ICR is being used to process insurance forms (that includes handwritten medical records) weeks faster and get patients their needed medical equipment.
More generally speaking, character recognition software is being used in the healthcare industry to digitize huge amounts of documents with patient data, such as:
The benefits are significant time savings, cost savings, and improved patient care.
Financial institutions such as banks and credit unions use ICR to streamline workflows through Optical Mark Recognition to get information off loan applications, forms, surveys and even checks.
They formerly were using human operators to tackle the painstaking job of manually entering the same information. Their customers (funeral homes) no longer have to spend as much time on paperwork, and how have more time to comfort grieving families.
Generally speaking, companies use ICR software to speed up insurance claims processing and policyholder information.
Law firms and legal departments use ICR solutions to extract handwritten information from forms contracts, case files, or handwritten forms.
The technology is also being used to automate the legal discovery phase, a process that was traditionally performed with painstaking manual work. It is usually performed by attorneys or highly paid legal assistants, so automation through ICR saves significant costs.
E-commerce and online-based businesses use ICR to collect electronic signatures and log them into databases to assist with know-your-customer documentation.
Every single industry deals with massive amounts of accounts payable data, in the form of invoices, receipts, bills of lading, etc.
ICR technology can be a vital component to achieving the highest rates of recognition accuracy, which drastically cuts down on human time needed to key the data into line-of-business systems.