What does a good biometric system consist of?

Welcome to our second post in our series on “The importance of optimizing all parts of a biometric system”. In this post, Peter Kollig, Senior Director of Engineering, System Architecture at IDEX Biometrics discusses what a good biometric system needs to consist of?

Most businesses and individuals now recognize the inadequacy of passwords as a means of securing systems. Biometrics is often the alternative choice, but if a system is complex or too difficult to use, people just won’t work with it, even if they’ve invested time or money in acquiring the technology.


The quality and performance of the biometric algorithm as well as the ease of the biometric smart card preparation prior to first use are paramount. Both are unique selling points of IDEX’s biometric solution.

Biometric templates have to be enrolled on the biometric smart card. With the IDEX enrolment sleeve, this can be achieved at the customer’s convenience at home. IDEX provides all parts of an elegant enrolment solution, including the sleeve, providing a mechanical guide for the user’s finger and LED signaling indicating where the user has to place the finger.

In addition to the sleeve technology, IDEX provides all relevant software modules on the Secure Element (SE) and Biometric Microcontroller Unit (B-MCU) Development tools are provided for large scale image collection, allowing customers to characterize their own systems.

Matcher characteristics

The embedded matcher is at the heart of the biometric system. It is the biometric verification algorithm that makes the difference between a biometric card that works well and a biometric card that is hard to use.

The matcher algorithm is constrained in many different ways:

  • The security requirements imposed by payment schemes mandate that biometric templates are stored securely in the non-volatile memory (NVM) of the SE.
  • The matching algorithm must be executed on the SE so that an attacker cannot manipulate any intermediate calculations or matching results that would spoof an incorrect payment authorization
  • The available computing power is typically limited to around 100 MHz and the processor cores in use are of Cortex-M class
  • SE typically have 12-16 Kbyte Static Random-Access Memory (SRAM) and 256-800 Kbyte NVM
  • MCU’s typically have 64-128 Kbyte SRAM and 256-512 Kbyte NVM
  • Capacitive FP sensors are typically less than 10mm for IDEX off-chip sensing, significantly smaller for silicon sensors

These key factors supporting a good biometric system should be at the forefront of any customer’s mind before deploying biometric technology. Using existing technologies and enhancing them with biometric systems that work seamlessly within the existing infrastructure can reduce the risk of data breaches. It can also streamline and improve ease of use, making the existing technologies more user-friendly. This makes for an easy-to-use and very safe biometric system.

The previous post in our series on “The importance of optimizing all parts of a biometric system” looked at what makes a successful biometric tool?