OCR or optical character recognition software

What is OCR software?

OCR stands for Optical Character Recognition. Although this technology was invented in the 1930s, it has been democratized over the last 20 years and greatly improved thanks to artificial intelligence (AI) and all the possibilities it opened up.

Concretely, this technological process is able to identify specific information from an image or a paper document. If an OCR is reliable enough, it can extract all the information from an initial image, such as texts, numbers, etc. Additionally, in some situations, OCRs can even preserve the style of the text (font, alignment, weights, underlining, etc.). Interest in OCR lies in its capacity to automatically recover and enter content in a computer program without human manual input, i.e., to convert paper files to digital format.

All your questions about OCR

All the answers to the most frequently asked questions about OCR!

In which sectors are OCR used?

Optical character recognition is used in a plethora of situations and activities, and it is difficult to draw up an exhaustive list.

• On a personal level, for example, it is not uncommon to use OCR software to extract the content of an invoice or a paper document to rework it in TXT or PDF format.
• But it is particularly in the professional world that the use of OCR is the most widespread. In the postal service, OCR automatically detects the addresses on the mail; in banks, it is used to sort cheques. The technology is also widely used in the legal, medical, and industrial fields.

In our case at Cegid Notilus, we make our OCR available to dematerialize our customers’ expense receipts.

How does OCR work?

Without going into too much detail, an OCR system will analyze and scan all the pixels of an image. When each character is highlighted, the OCR software will then try to understand if it means a letter, a number, or a symbol. It will then transcribe it into a computer program.

For personal use

There are many free OCR tools. Most of them are even available online and do not require installation, so you are spoilt for choice!

In the professional world

A company wishing to develop its own OCR can start from an open-source base or use the technology offered by Amazon or Google, which is easier to use and less laborious.

OCR + artificial intelligence = intelligent scanning

By combining Machine learning and Deep learning technologies to an OCR, we greatly improve the latter’s accuracy. In fact, the OCR will even progress automatically from day to day. Artificial intelligence algorithms will continuously challenge the OCR software by mimicking human reasoning, and train it over various databases, to improve its prediction rate.

OCR in the context of expense reports and supplier invoices

In France, with the promulgation of the decree of March 22, 2017, and May 23, 2019, businesses are now free to dematerialize their supplier invoices and digitize their expense receipts.

If some studies estimate that it takes on average 20 minutes for an employee to write up an expense report, imagine how much time could be saved for a company with frequent professional travels, and its accounting department.

To digitally store receipts of corporate expenses, electronic archiving with probative value is a must and a legal obligation. All dematerialized receipts are kept safe in an encrypted strongbox. Notilus uses Arkhineo technology to guarantee the probative value of users’ receipts and make them available for 10 years. These documents are valid in case of tax or French URSSAF audits.

Thanks to our professional OCR software, the entry of expenses in the application is entirely automated and the employee only has to verify that all information is correct. Dematerialization of expense reports brings better control, reduction of costs, and a high level of security.

Cegid Notilus OCR to simplify your expense reports

To put an end to the time-consuming and tedious task of filing expense reports, Notilus’ artificial intelligence team has developed our very own OCR technology, which automatically recognizes the content of expense receipts and other proof of payment documentation.

To use it, one only needs to open the Notilus mobile application on their smartphone and take a picture of the receipt. Et voila! Language, type of expense, date, amount, VAT rate, currency, country, etc. are instantaneously recognized and the fields on the automated expense reports are filled in automatically.

It takes an average of 2 to 3 seconds for the Cegid Notilus OCR to transcribe content. Behind this remarkable computational speed is a database with some 7.9 million expense receipts that are continuously scanned by the OCR software and its AI algorithms. In addition, our OCR offers high reliability, with an average recognition rate of 95%. But action speaks louder than words, please don’t hesitate to contact us if you wish to try it out yourself!