Joshi brings more than 20 years of product leadership experience to his new role at Veriff from his previous positions at Amazon, Dropbox, Facebook (Meta), Microsoft and Yahoo. In his most recent role, he worked as a Chief Product Officer at FarEye, where he was responsible for accelerating the company’s global product expansion.

“Having experience working with both large, established companies as well as scale-ups, Suvrat will play a vital role in Veriff’s future product offerings as our customer demands continue to change and we are expanding into new industries,” said Janer Gorohhov, Chief Product Officer and co-founder of Veriff. “The demand for identity verification (IDV) is developing at a rapid pace, and we’re honored to have Suvrat and his proven expertise at the helm of Veriff’s product engine as we continue to navigate and expand within the market.”

“The coupling of digital and physical worlds through the identity lens has the potential to transform existing and emerging industries, and Veriff is at the very center of this intersection,” said Joshi. “It’s an exciting time to be joining Veriff, and after already witnessing some of the great value the company brings to multiple industries, I am thrilled to have the opportunity to help the company continue to make an impact and shape the identity verification space.”

Joshi holds a Bachelor in Engineering from Pune University, Masters in Engineering from Rensselaer Polytechnic Institute (RPI) and a MBA Degree from UC Berkeley.

We are thrilled to announce Suvrat Joshi joining Veriff as Senior Vice President of Product. As a part of Veriff’s senior leadership team, Joshi will be responsible for leading the company's product strategy and spearheading Veriff’s next generation of product offerings to continue to support the evolving needs of Veriff’s existing and new global customers.

A blog post announcing Suvrat Joshi as Veriff's new Senior Vice President of Product

Suvrat Joshi joins Veriff as Senior Vice President of Product

Optical character recognition (OCR) technology can save your business time, money, and effort. This is because OCR uses automated data extraction and storage capabilities that can improve the accuracy of data input and create quicker file digitization processes. 

OCR is sometimes referred to as text recognition. This is because OCR programs are commonly used to extract and repurpose data from scanned documents, camera images, and image-only PDFs.

OCR enables users to access and edit the original content. But, OCR technology can be used for much more than this. For example, it can also be used to eliminate the need for manual data entry, recognize information on passports and traffic signs, and to test CAPTCHA anti-bot systems.

Early forms of optical character recognition involved creating reading devices for the blind. For example, in 1914, Emanuel Goldberg developed a machine that read characters and converted them into standard telegraph code.

Following this, in the 1920s and 30s, Emanuel Goldberg developed a statistical machine for searching microfilm archives using an optical code recognition system. In 1931, he was granted a US patent number for the invention. This patent was later acquired by IBM.

A more modern form of optical character recognition was first introduced by Ray Kurzweil in 1974. His company, Kurzweil Computer Products, Inc., created an omni-font OCR product that could recognize text that had been printed in (almost) any font. Once he’d designed the product, Kurzweil decided that the best application of this technology would be a machine-learning device for the blind. As a result, the company also created a reading machine. This device was used to read text aloud in a text-to-speech format.

In 1976, the finished product was unveiled during a widely reported news conference headed by Kurzweil and the leaders of the National Federation of the Blind. Only two years later, Kurzweil began to sell a commercial version of the computer program.

After experiencing initial successes, Kurzweil Computer Products, Inc., caught the attention of XEROX, who purchased the company in 1980 and began to commercialize paper-to-computer text conversion.

Partially due to this, OCR technology gained mainstream popularity in the early 1990s, when it was primarily used to digitize historical documents and newspapers. Before OCR technology first became available, the only way to digitally format these documents and newspapers was to manually retype the text. However, this process was incredibly time-consuming and led to a number of inaccuracies and typing errors. Although these early versions of OCR technology were not perfect, they were far more accurate than humans and saved a great amount of time.

Fast forward to the year 2000 and OCR was made available online as a service in a cloud computing environment and in mobile applications. Since, OCR technology has been made available in internet connected mobile device applications that extract text captured using the device’s camera. Now, various commercial and open source OCR systems are available for most common writing systems, including Latin, Arabic, Hebrew, Tamil, Chinese, Japanese, and Korean characters.

Today’s OCR technology is far more advanced than early versions of the technology. This is because today’s advanced solutions have the ability to deliver near-perfect OCR accuracy and can be used to automate complex document-processing workflows. But, not only are the solutions much more advanced and accurate, but they’re also applicable to a far greater number of use cases. Let’s take a look at some of these in greater detail.

To the modern user, OCR technology may seem like an abstract and technical concept. However, the opposite is the case and many of us encounter OCR technology in our lives on a regular basis.

On top of this, OCR services are also widely available to the public. For example, Google Cloud Vision OCR can be used to scan and store documents on any smartphone. You can even try it for free on Google Cloud.

With this in mind, let’s take a detailed look at some of the ways OCR technology is used.

OCR is primarily used for data entry purposes. It can input data from printed paper records and business documents including passports, invoices, bank statements, and receipts.

OCR is also a common way to digitize printed text. Following OCR, documents can be electronically edited, searched, stored more compactly, displayed online, and used in machine processes such as cognitive computing and machine translation.

OCR can also be used for automatic number plate recognition. In these instances, OCR can read a vehicle’s registration plates and identify a vehicle’s location. To do this, it can utilize existing closed-circuit television and road-rule enforcement cameras.

In these scenarios, automatic number plate recognition is used by police forces to check if a vehicle is registered or licensed. It is also used for electronic toll collection on pay-per-use roads, for parking enforcement purposes, and as a method of cataloging traffic movements. 

In airports, OCR technology is used for passport recognition purposes. At borders, it can be used to extract information from a passport that is placed on a scanner. This information is then scanned against third-party databases to ensure the passport is authentic, the individual is real, and the passenger is allowed to fly.

OCR is also often used to extract key pieces of information from documents. In lengthy documents (such as insurance documentation), it can be difficult for people to find the information they need at a glance. OCR makes it possible to find this information in an instant. In a business capacity, this leads to huge time savings and reduced admin time.

With traffic sign recognition technology, vehicles can recognize the traffic signs on the road ahead and alert the driver. Commonly, this technology utilizes OCR and then tells the driver about upcoming road safety features, such as a speed limit or whether they’re approaching a sharp turn or a road with restricted access.

OCR technology can also be used to extract business card information. Once information has been extracted from the business card, it can be placed into a contact list. This saves time, makes data processing easier, and creates shareable leads for sales teams.

OCR can be used to make text visualizations of printed documents. For example, OCR can be used to scan books and digitize them, turning them into eBooks. It can also turn magazines into digital media.

This technique is most popularly associated with Project Gutenberg, a volunteer effort to digitize and archive cultural works. The world’s oldest digital library, Project Gutenberg has digitized more than 60,000 documents and made them freely available in the public domain. All digitized files can be accessed under an open format layout, which is available on almost any computer.

Converting handwritten text (pen computing)

Pen computing (the act of a computer user-interface using a pen or stylus and tablet instead of a keyboard or a mouse) is reliant on OCR technology. This is because the technology has the capacity to convert handwriting in real time.

Although CAPTCHA anti-bot systems have been specifically designed to beat bots, there’s some evidence that OCR can still bypass poor systems. For this reason, OCR technology is still widely used to test the robustness of CAPTCHA systems and find any weaknesses before the systems are rolled out.

OCR technology is commonly used to write instructions based on CAD images. This means that instructions can be written in real time as a product is designed.

Finally, OCR technology is also regularly used to make scanned documents searchable. This is made possible because OCR technology can convert printed documents to searchable PDFs.

Types of optical character recognition technology

Optical character recognition is generally an offline process that analyzes static documents. However, some cloud-based services do provide an online OCR API service.

Different types of OCR are available, and the four main types are:

This form of OCR targets typewritten text, one glyph (purposeful mark) or character at a time.

This type of OCR also targets typewritten text, but does so one word at a time. This style of OCR is particularly useful for languages that use a space as a word divider.

By contrast, intelligent character recognition (ICR) targets handwritten printscript or cursive text one glyph or character at a time. This process usually involves machine learning, which improves accuracy.

Similarly, intelligent word recognition (IWR) also targets handwritten printscript or cursive text, but does so one word at a time. This type of OCR is used most regularly for languages where glyphs are not separated in cursive script.

OCR techniques vary depending on the use case and the type of OCR technology. However, most forms of OCR include techniques such as pre-processing, text recognition, and post-processing. Some also involve application-specific optimization. Let’s look at each of these techniques in greater detail.

OCR software often pre-processes images. In doing so, the software improves the chances of successful recognition. A number of pre-processing techniques are used widely, but the most common options include:

Deskewing the document to ensure proper alignment

Despeckling the document to remove positive and negative spots

Converting an image from color or grayscale to black-and-white in order to improve the quality of the character recognition

Line and word detection, which establishes baseline for word and character shapes

Layout analysis, which involves the identification of columns, paragraphs, and captions

When it comes to text recognition, there are two basic types of core OCR algorithm:

Pattern matching. Here, OCR programs are fed examples of text in various fonts and formats. These are then used to compare and recognize characters in the scanned document

Feature extraction. Here, OCR programs apply rules regarding the features of a specific letter or number to recognize characters in the scanned document. Features could include the number of angled lines, crossed lines, or curves in a character for comparison

Some pieces of software use a two-pass approach to OCR. Here, the second pass is known as ‘adaptive recognition’ and uses the letter shapes recognized with high confidence on the first pass to improve recognition of the remaining letters on the second pass. Alternatively, some modern pieces of OCR software use neural networks, which are trained to recognize whole lines of text instead of focusing on single characters.

But, the technology surrounding OCR is constantly evolving and new techniques are continually emerging. For example, iterative OCR can now automatically crop a document into sections based on the page layout. OCR is performed on the sections individually using variable character confidence level thresholds to maximize page-level OCR accuracy.

Post-processing can improve the accuracy of OCR. Depending on the complexity of the document, the post-processing phase may simply involve a human proof-reading and checking the document before circulating it.

The accuracy of OCR can be increased if the output is constrained by a lexicon, such as all the words in the English language or a technical lexicon for a certain field. For this reason, some pieces of OCR software use a dictionary to influence the character segmentation step and improve accuracy.

Similarly, by conducting a near-neighbor analysis during post-processing, businesses can make use of co-occurrence frequencies to correct errors. In certain industries and niches, algorithms such as the Levenshtein Distance algorithm are used in OCR post-processing to further improve results.

The advancement of OCR technologies and techniques means that providers of optical character recognition systems have tweaked their offerings and ensured that these systems can now deal with specific types of input. This means that these systems can now take into account business rules, standard expressions, or rich information contained in color images.

Known as application oriented OCR or customized OCR, this form of optical character recognition has been applied to the scanning of license plates, invoices, screenshots, ID cards, driver licenses, and automobile manufacturing.

Prevent customers being the victim of identity verification scams with Veriff - Request a demo

If your business needs to verify customer identity documents, then our ID verification software is perfect.

Using OCR technology, it can verify identity documents in real time and detect fake and tampered documents. It extracts barcode, MRZ, and NFC data and compares this against identity documents. Using webhooks, you can instantly receive primary data, such as the names, birthdates, document numbers, and other information on the document. It covers more than 10,000 IDs from more than 190 countries and in more than 45 languages.

If you’re looking to verify customers and ensure data accuracy, then enlist the help of our biometric authentication solution. This way, you can confirm that any returning user is exactly who they’re claiming to be.

With the help of our solution, you can secure customer accounts and take a step beyond passwords and one-time passcodes. You can also make the authentication process simpler and more secure. Users can be authenticated in only a second with 99.99% accuracy.

If you’d like to learn more about how any of our solutions can help your business, contact our experts and arrange a free demo today.

Today’s OCR technology is far more advanced than early versions of the technology. This is because today’s advanced solutions have the ability to deliver near-perfect OCR accuracy and can be used to automate complex document-processing workflows.

A blog about optical character recognition, including the history of the technology, examples of when the technology is used, and types of OCR.

What is optical character recognition? (OCR)

Find everything you need to know about optical and biometric passports in this comprehensive guide.

Whether you own an optical or biometric passport, you’ll be able to pass through border control without any issues (as long as your passport is still in date). But while both of these documents accomplish the same mission and both forms of identity document are perfectly valid, there are some differences between the two. Here, we’ll cover whether optical or biometric passports are the best option and why.

Optical passports are machine-readable passports. These documents contain the personal information of the passport holder in two lines of numbers, letters, and arrows at the base of the main passport page.

Originally introduced in the 1980s, optical passports quickly became popular because they allowed immigration officers to process the details of arriving passengers quickly. Plus, optical passports were found to be more accurate than manually-read passports. Added to this, data entry was also faster and it was much easier to match this data to information on immigration databases and watchlists.

However, although optical passports were incredibly popular in the 1980s and 1990s, most have now been replaced by biometric passports in order to increase and reinforce the security of personal information.

First introduced in Malaysia in 1998, biometric passports (which are also known as e-passports or digital passports) include embedded electronic microprocessor chips that hold personal information, as well as the owner’s photograph.

Biometric passports are still traditional paper passports. However, they provide a number of security advantages over optical passports. For example, the chip and the data held within a biometric passport is much more difficult to replicate than information on a printed page. This makes it much tougher for a fraudster to either create a fake passport or assume someone else’s identity.

On top of this, biometric passports are also popular because they make the border experience far simpler for the passport holder. With a biometric passport, a passport holder can use the e-passport gate when they reach border control. A machine scans the passport and matches it with the data on the holder’s face using facial recognition technology. This speeds up the process and reduces the likelihood of lengthy queues.

Optical passports contain encoded information in optical character recognition format. This information can then be read by a computer with a camera or suitable piece of software at border control.

Information encoded includes details about the traveler, including their full name, date of birth, and passport number. After scanning the information, checking it against databases, and ensuring the person in the picture is the person in front of them, the border control agent can allow the traveler to enter the country.

By contrast, biometric passports contain an embedded electronic microprocessor chip. This chip contains biometric information that is used to authenticate the identity of the passport holder. As well as being stored in the chip, important information is also printed on the name page of the passport.

One of the key differences between the two documents is that biometric passports do not need to be checked by a border agent. Instead, the traveler can walk to a dedicated biometric passport area at border control and place their passport on a scanner and contactless chip reader.

Following this, a passport data server will then check the validity of the passport and the data it holds. It will then also check that the person standing in front of the camera is the person pictured on the passport by comparing biometric data. If a match is confirmed, a barrier will open and the traveler will be allowed to proceed into the country.

Most countries around the world no longer issue optical passports and instead only issue biometric passports. However, in the majority of countries, optical passports remain valid as long as they’re in date. This means that regardless of whether you own an optical or biometric passport, you can pass through border control.

However, one prominent exception to this is the United States of America. Any foreigner traveling to the U.S. who wishes to enter visa-free under the Visa Waiver Program (VWP) must possess a machine-readable passport that complies with international standards. If issued on or after October 26, 2006, this passport must be biometric in order for the individual to enter the U.S. visa-free under the VWP.

Biometric passports are now issued instead of optical passports because they provide a number of benefits for travelers and border police alike. For example, e-passports help to:

Make it incredibly difficult for an existing document to be altered

Make it incredibly difficult for a fake passport to be created  

Ensure international co-operation with passport security

Make sure passengers are prevented from encountering delays at borders

In addition to this, there are multiple layers of security in the e-passport that prevent fraud. Common security features contained within biometric passports include:

Plus, rather than relying on a simple photo of the individual in question that is then checked by the eye of a border official, biometric passports harness the power of biometric data (as the name suggests). This in itself provides greater security because biometrics are unique, measurable characteristics that can be used to verify the claimed identity of an individual. For this reason, biometrics are widely considered to be one of the most secure types of authentication methods in use today.

Examples of the types of biometric data that can be included within a passport are facial images, fingerprint scans, and iris patterns. By combining different forms of biometric identifiers, passport issuers can create an unrivaled level of security and protection against counterfeit and fraudulent identification documents.

From the above information, you may not be surprised to learn that biometric passports are the most secure passport type available.

However, it’s important to remember that passports issued by different countries during different periods of time can also be different. In the UK, for example, passports issued after 2010 are different to the initial e-passports issued in 2006.

That said, the International Civil Aviation Organization publishes standards for both machine-readable passports and biometric passports to ensure that passports issued by different countries meet certain criteria and standards.

When measured by mobility score (the number of countries that allow the holder of that passport to enter for general tourism visa-free, visa-on-arrival, eTA, or eVisa issued within 3 days), the strongest passports belong to:

Germany, Sweden, Finland, Luxembourg, Spain, France, Italy, Netherlands, Austria, Switzerland, and South Korea (170)

Denmark, Norway, Belgium, Portugal, Poland, Japan, Ireland, United States, and New Zealand (169)

Malta, Czech Republic, Greece, Hungary, United Kingdom, and Australia (168)

Singapore, Lithuania, Slovakia, and Canada (167)

Although they’re a relatively modern invention, biometric passports are now used by more than 150 countries worldwide. Malaysia was the first country to issue biometric passports to citizens, back in 1998. Ever since, countries have quickly adopted these digital passports. Within the first decade, an additional 60 countries began to issue them to citizens.

Biometric passports were first issued in the United Kingdom and the United States of America in 2006. Since, both countries have stopped issuing non-biometric passports.

In total, it’s estimated that more than 1 billion people use e-passports to cross borders around the world.

See how Veriff’s biometric authentication solutions can help you - Book a demo

If you run a business that needs to verify the identity of its users and meet stringent AML and KYC regulations, then our biometric authentication solution can help you achieve your aims.

It harnesses the power of biometric data and confirms that a returning user is exactly who they’re claiming to be. It’s far more secure than using passwords and one-time passcodes and is perfect for securing user accounts.

As soon as the user provides our software with a selfie, it’s checked for liveness and realness. Then, the selfie is compared to a previously identified face and identity. If a match is found, a user can access their account.

Any user’s identity can be authenticated in less than a second. Plus, the solution is 99.99% accurate and 99% of users are authenticated on their first try.

Interested in learning more about how our biometric authentication solution can help your business? Book a demo with our experts today. 

Although optical passports were incredibly popular in the 1980s and 1990s, most have now been replaced by biometric passports in order to increase and reinforce the security of personal information.

A blog about optical vs biometric passports, including the history of optical and biometric passports, how each type of passport works, and countries that employ biometric passports today.

Supported languages

Want to learn more?

We're certified

Want to learn more?

Suvrat Joshi joins Veriff as Senior Vice President of Product

What is optical character recognition? (OCR)

Optical vs biometric passports

Products

Solutions

Resources

About