We’ve noticed that people often believe Regula’s Information Reference System (IRS) and the identity document template database built into the Regula Document Reader SDK to be the same.
This post explains the differences between these two databases and outlines the specific use cases for each.
|Information reference system database
|Regula Document Reader SDK database
A collection of detailed reference document images with their descriptions
A set of software algorithms trained to process specific document types used by Regula Document Reader SDK
Manual document examination
Automated document verification
Expert-level in-depth comparison performed by:
Document templates (software algorithms for recognizing a document, its structure, and data).
Level of detail
What’s the IRS?
Regula’s IRS, or information reference system, is the world’s most detailed database of secure document samples for expert document examination. Depending on the version, it may include different categories of items:
passport database along with visas,
…or all of them at once!
The idea behind the IRS is simple: no inspector can be expected to memorize every ID document in the world. The IRS enables them to find out what a certain document looks like, covering everything from front and back covers to internal pages, what security features it should have, and crucial background information about the document.
Each document within the IRS is meticulously marked and labeled, allowing users to observe various security features in detail. For dynamic security features that alter their appearance with movement, animated GIFs are included to represent these changes visually.
Today’s version of the IRS, the most complete yet, comprises over 11,500 documents, banknotes, and coins, totaling over 300,000 images.
Users have the flexibility to select the depth of the database. The basic version includes images of documents in white light, with only data pages captured under UV and IR and front covers. In contrast, the advanced version provides all document pages under white and UV (including hard UV), IR, and anti-Stokes, along with magnetic pattern visualizations.
It’s important to note that some versions of the IRS can only be accessed by law enforcement agencies and authorized forensic laboratories.
There’s also Document Builder, an additional IRS module that allows users to create their own reference database. For example, some countries use it to gather detailed information about all the forged and counterfeit ID documents they detect to prevent similar attempts in the future.
What’s an identity document template database?
Regula’s identity document template database is used for automated document checks, such as the ones that take place during client onboarding in bank apps or traveler ID checks in airports.
For a better understanding of what it is and how it works, it’s crucial to define what a “template” is.
A document template is a software algorithm designed for the machine-based processing of a specific document. In contrast to the IRS, the document template database doesn’t include actual document images. Instead, it comprises various parameters and binary data utilized by the SDK.
Each template contains encoded information about the physical locations of various fields and elements within the document, their properties, possible variations in filling, and tolerated deviations due to production specifics and wear and tear.
The template’s purpose is to enable software to identify, read, verify, and authenticate the document when submitted to the system. Once the software identifies the document type, it retrieves general document and personal data, as well as its format, language, and other characteristics/attributes, including main and secondary portraits, based on the established template. Then, it moves on to automatic lexical analysis, data cross-validation, and authenticity check.
Importantly, a template is not equal to the document type itself.
There can be situations when two documents have three templates. For example, Japanese driver’s licenses may differ on the front side, but have the same back layout. As a result, these two driver's license types would have two different templates for the front sides while sharing a single template for the reverse side.
The key differences between the IRS and Regula Document Reader SDK databases
Purpose. In their basic configurations, the IRS is tailored for manual comparison and expert examination, whereas the SDK identity document template database is designed to facilitate automated document verification.
If an automated verification scenario requires document authenticity check (for example, e-gates in airports), then there should be a capability to capture documents in UV and IR light. This can be obtained using specialized document readers, which are designed exactly for this purpose.
Production. The speed of adding both a document template into the database and a specimen into the IRS mainly depends on whether a particular document type contains any security features that are totally new to Regula or not.
Basically, to make a new template for the SDK, it’s enough to put a document in a document reader or take photos of it in the most urgent cases. Together with marking up, the production of a template may take one working day.
This process is iterative, though. Say we might use some photos of a brand-new document to rapidly create a new template and refine it once we have more specimens.
If there’s a new security feature, forensic experts thoroughly examine it using professional devices. Then the team creates and tests new verification algorithms and rolls out an update for all documents with this feature.
The production of a new reference sample document for the IRS requires about two weeks in a lab for our forensic experts to examine it in all available light sources, capture all the security features, and describe them in detail.
Quality. To capture documents and their security features for the IRS, Regula’s experts use special professional equipment, for example, the Regula 4308. This dual-video spectral comparator has 30 types of light sources, over 20 light filters, and optical magnification capabilities of up to 320x. Thanks to these capabilities, they get super-detailed reference images that serve as a source of truth when determining the authenticity of a questioned document.
In real-world shooting conditions, the visual quality of submitted images is usually lower. This fact is taken into account when creating the document templates.
How these two databases fit into real-world workflows
To illustrate the practical application of these databases, let’s consider a scenario at an international airport border control post.
After an officer visually inspects a passport and its holder and completes a visual check of the basic security features, they place the passport in a document reader. These devices allow officers to instantly get results of automated document verification. Here, the SDK identity document template database comes into play.
Once the passport images are captured, the software queries the template database to identify the document type, determines the data to be extracted and analyzed, specifies the further verification process, and then actually verifies it.
This is the so-called first line of border control.
If the border officer sees red flags from the software or has any doubts about the passport's authenticity, it’s escalated to the second line of control. At this stage, senior officers will manually examine the document using more advanced forensic devices and the IRS.
💡The same process can be employed by any industry where reliable identity verification is essential, be it Banking or Fintech. This way, businesses can ensure the same level of security as border controls.
For certain scenarios, the IRS can be integrated into the software used for automated document reading as an additional module. This hybrid approach empowers inspectors to access reference images of a document directly within the program interface.
- The IRS is for manual expert examination, and the SDK document template database is for automated document verification.
- Documents presented in the IRS are reference images captured in a perfect laboratory environment.
- Some versions of the IRS can only be accessed by law enforcement agencies and authorized forensic laboratories.
- Regula Document Reader SDK enables instant data extraction and verification of the consistency of the data and its format. It can also be used for automated document authenticity checks, using specialized devices for examining in UV and IR light.
- The SDK document template database doesn’t consist of actual individual documents. Each item is an algorithm that “knows” how to handle a certain document, from detecting its type to verifying security features and data fields.