As you are reading those words, you are able to understand and process them just by using your eyes and brain. You recognize not only whole words and sentences but also letters, numbers and punctuation. This is basically optical character recognition. Is there a way for computers to do exactly the same with paper documents, handwriting and other “unreadable” files? Yes, and it is called OCR technology.
In this article, we will explain how OCR software works, where optical character readers are mostly used, and how your company can benefit from introducing OCR to document processing and content management.
What is OCR and how does it work?
Optical character recognition (OCR) software is a mechanical or electronic method of converting handwritten, typed, scanned, photographed, faxed, printed (both paper and digital) text files into machine-encoded, editable and searchable documents in nearly any format.
To put it simply, computers detect pictures, not text itself. A printed page is just a bunch of meaningless pixels, so for computers there is no difference between a picture of the New York skyline, your child’s Paint scribbles, a scan of the Mona Lisa, or an invoice for a new laptop. They cannot “read” the invoice any more than they can read the other examples. Getting the text in order to turn a jpeg into a doc file quickly and accurately is the main task of OCR software.
The idea for human-machine communication and information exchange is not new. The beginnings of OCR technology can be traced back to the 19th century, when the first devices to help the blind were designed. In 1912, Emmanuel Goldberg was granted a patent for a machine that was able to read characters and convert them into telegraph code. Later, the telegraphic message was sent over wires with no human intervention. The first OCR-like system was invented in 1929 by Gustav Tauschek. His machine processed text via templates in front of a photodetector. It was able to recognize characters by comparing them with the font on the template. Unfortunately, a tiny difference in shape rendered a text unrecognizable. OCR software has progressed a great deal since then, and brand new OCR readers are equipped with thousands of fonts and hundreds of dictionaries to give a most likely word match in case of error.
Nowadays, OCR technology is mostly used in business to handle paperwork and make business content smarter to create workflows for document exchange. OCR software comes in different forms, including personal smartphone scanning devices, desktop applications, advanced business platforms and server-based solutions that are capable of converting millions of files with no human interaction. So, if your company has ECM or EDI software, you might consider adding an OCR reader as additional feature.
How does OCR “read”?
There are a few different ways in which OCR software processes texts. These are pattern recognition (recognizing characters in their entirety), feature detection (recognizing individual elements of each character – lines, strokes, angles, etc.), intelligent character recognition (for handwriting), optical mark recognition (recognizing answers in tests and surveys, where respondents have to fill in a blank square using a pencil) and optical barcode recognition (recognizing a product and its features by scanning a barcode). By combining all of these methods, it is possible to read text from any given document, including orders, invoices, letters, contracts, forms, reports, personal questionnaires, tables of content and many others. Considering this is an article about text recognition, let’s focus on the first three methods.
Using this method, the computer recognizes the entire character and matches it to the characters in the stored OCR software. This started in the 1960s, when a special monospace font (OCR-A) was created for filling in official bank checks. Every letter was easily distinguishable from the others. Both the printers and scanners used only that font. It was an effortless OCR system because of this single, standardized font. Later, more common fonts were added. Unfortunately, there are now so many different font types that pattern recognition alone might be insufficient to recognize all characters, let alone handwriting.
Where pattern recognition fails, feature detection is there to fill in the blanks. Instead of recognizing the letter as a whole, this OCR reader detects individual components such as horizontal, vertical and diagonal lines, angles, points where lines meet, closed loops, intersections, diacritical marks. You do not need a database of fonts, as the rules of writing a letter or digit apply to all designs. Most modern OCR software that can recognize print in any font operates using feature detection.
Intelligent character recognition
This decodes handwriting and cursive script one character at a time and involves a machine learning engine. It is easy for humans to read handwriting (even if we all know someone who seems to write in hieroglyphs), for we understand the language, context and intent of the writer. It is extremely difficult for computers to do the same, but with a little help it is not impossible. OCR-friendly forms often have empty boxes to fill in with capital letters and dark ink, so the computer has to identify only a relatively small amount of information.
You have probably helped develop OCR technology yourself. CAPTCHA puzzles are used to stop bots and spammers from messing with email systems, forums, message boards, ticket offices and so on. But it has a second purpose: millions of humans help computers recognize words that are too complicated for them to read. It might be a scanned, barely visible text from an old book that is being digitized, but everything helps.
Where is OCR useful?
OCR technology is widely used in many different industries. Apart from giving you business advantages, it can actually help visually impaired people to read, and historians to preserve culture. Below we present a few examples of beneficial uses of optical character recognition:
Helping blind and visually impaired people
In the 1970s, Ray Kurzweil started a company Kurzweil Computer Products Inc. This firm delivered the world’s first omni-font OCR system. This OCR computer software was able to recognize the most popular printed fonts, and read the text in books and magazines out loud to blind people. In other words, the text was not only decoded but also transferred to speech synthesizer tools (the most famous computerized voice was Stephen Hawking’s of course). The machine was presented at a special conference organized by the National Federation of the Blind.
Sorting letters in post offices
The United States Postal Service (USPS) processes and delivers 493.4 million pieces of mail each day. That's an average of 20.6 million each hour, 342,638 per minute and 5,711 every second. Who is responsible for sorting all this correspondence? Fortunately, OCR readers are on hand to help the human employees. These devices are able to decode not only computer generated labels, but also handwritten addresses and ZIP codes. Imagine the delays if all those pieces of mail were handled manually.
Sorting documents in law firms and courts
The most complex legal cases and investigations are bottomless pits of documentation. Lawyers have to dig painstakingly through every page to make sure they do not omit any important detail. OCR software helps them digitize all necessary files and search for keywords, dates, names, etc. This means they do not have to hire more people to do the job manually and risk oversight due to fatigue or human error.
Preserving historical and cultural texts
Libraries, historical organizations and NGOs (such as Project Gutenberg) archive books, manuscripts, documents and memoirs, and process them from paper form to digital files. Manual retyping of that amount of text would be impossible. Thanks to OCR systems, this task is not only doable, but the process of preserving world’s heritage is simplified.
With OCR software, legal documents such as passports, drivers’ licenses, passports, car number plates, insurance certificates and more can be scanned quickly by offices, police, airports and other institutions that process our personal information. OCR technology helps us reduce human error and obtain accurate data.
Processing invoices, orders and other documents
A lot of organizations convert paper invoices and contracts to digital text in order to use data for financial reports, payments, and document exchange. There are multiple document sources and formats, including handwriting on paper, PDFs, jpgs, etc. It is difficult for employees to enter all such data manually into a company’s systems, which ultimately results in wasted time, a never-ending pile of paperwork and far too many mistakes. In this respect, OCR software is a blessing for businesses. Research shows that companies using OCR for automated data entry from invoices process 25-60% more invoices than companies using manual data entry. What is more, OCR technology makes it possible to register invoices in the company’s electronic archive automatically, regardless of whether they arrive by traditional post, email, fax or EDI platform.
The benefits of OCR for your business
In a survey prepared by Paystream Advisors, respondents were asked to rate the “pain” associated with certain invoice processing activities. On a scale of 1 to 5 (where 5 was the biggest inconvenience), manual data entry had the highest average pain score (3.27), followed by resolving errors (3.17). This clearly shows that employees dealing with documents absolutely hate the repetitive typing tasks and the human errors that accompany them. OCR software helps reduce both, and brings many benefits to your company. A few advantages are listed below:
Automatic digitization of data happens much faster than manual processing. Files are archived in minutes instead of hours, and retrieving information or an archived document happens in the blink of an eye.
Better work management
Give your employees the appropriate tools and you will notice improved efficiency. They will not spend time on repetitive, boring tasks, and focus more on previously neglected duties and more productive work.
Reduction of costs
Companies spend $20 in labor to file a document, $120 in labor to find a misfiled document, and $220 in labor to reproduce a lost document. In other words, an OCR system which reduces manual labor and human errors saves you money. Investing in OCR software will eventually pay off.
Fast research tool
It is much faster to search for a particular piece of information when you have a digital, searchable file, than to go through endless paper documents page by page. You staff will thank you for improving their working environment by reducing the amount of paper documentation.
Improved business processes
Manual data processing is time-consuming and does not generate any direct value for your business. It may cause data errors, delays, misunderstandings, wrong orders, warehouse stock problems and a slow business process cycle.
Security of documents
Invoices, contracts, clients’ personal data and other important information must be secure 24/7. Any security breaches undermine your reputation and can result in lost clients. OCR is often part of a broader solution that also archives your crucial data securely.
No more storage problems
Massive file cabinets located in every office and containing tons of folders and boxes take up a lot of space. The stream of documents seems to grow until you have absolutely nowhere left to store paper documentation. A digital archive stores documents processed by OCR, making it easier to find a given file.
Improved customer service
It is not a good business practice to keep a customer waiting, especially when an upset client calls your consultant and they need to access their documents instantly. Thanks to OCR, it takes seconds to collect and retrieve client information, and if you speed up and simplify customer service, you will improve overall customer experience as well.
Good for the environment
This is not only about the tons of paper being used. Consider also wrapping foils, tape, fuel used for shipping documents, ink, and so on. Eco-friendliness is valued by customers as well, so that is an additional asset for your business.
OCR in Comarch Enterprise Content Management (ECM)
Optical character recognition is one of the features of Comarch Document Management. It is based on the Enterprise Content Management (ECM) system, which allows a synergistic approach to documents, data and processes. The company has the ability to register documents regardless of format, to archive them securely, and to access them quickly, particularly in the context of tasks arising from business processes.
A properly implemented OCR system in Comarch ECM allows:
- Efficient “reading” of data from documents in various formats, including scanned documents, faxes, e-mails and their attachments, Office suite documents, photographs, etc.
- Intelligent data recognition and auto-configuration of the OCR module based on verified information provided by the corrector (without preconfigured templates)
- Full search of keywords in documents of any format
- Automatic recognition of documents and their separation during processing
- Automatic processes for specific types of documents (e.g. archiving and forwarding)
Thanks to OCR software, the costs related to the human factor, automated work in companies and public institutions, and quick access to key data from various sources (graphic documents, PDFs, paper documents, e-mails, faxes, etc.) are minimized. Additionally, integration with various external systems (e.g. EDI platforms) or internal client systems (ERP), makes Comarch ECM a single point of access for all of your company’s business information.
Optical character recognition is also very helpful for visually impaired people in their everyday lives, and for preserving cultural and historical heritage. While OCR technology is still being perfected, we predict that it will become widespread and popular.
Learn more about: