Sony Global - Technology

Technology

Activity

R&D

Others

mofiria(Finger Vein Authentication Technology)

The proliferation of network products and services has led to an increased need for personal identification and interface systems which are both user-friendly and highly advanced in their ability to protect personal information. The "mofiria" finger vein authentication system developed by Sony is easily mounted on personal computers and mobile devices, including mobile phones. Thanks to its unique optical mechanism and algorithm, it offers high-speed response in a highly-compact package. The "mofiria" system doesn't require the finger be held in a certain position because finger position is automatically accounted for and simultaneously corrected as the vein pattern is quickly and accurately reproduced from the captured image.

What is Finger Vein Authentication?
Compact Design Thanks to Reflective Dispersion Technology
Four Processing Algorithms for Finger Vein Patterns
Data Compression Reduces Data Volume to Approximately One-Tenth
Automatic Finger Position Correction, High Speed Processing
Looking Ahead

What is Finger Vein Authentication?

Finger vein authentication is a method for identifying individuals according to vein patterns in their fingers. Vein patterns vary from individual to individual, as well as finger to finger, even on the same hand. These patterns do not change over time. Because authentication is based on veins located within the human body, falsification is extremely difficult, and authentication accuracy is superior to other biometric methods.

Compact Design Thanks to Reflective Dispersion Technology

Sony's "mofiria" employs a unique reflective dispersion method. Infrared light emitted by an LED is directed at the veins, causing light to be dispersed inside the finger. A CMOS sensor captures this light and efficiently turns it into an image. With the reflective dispersion method, LED light is diagonally projected at the veins to produce images. This ensures that the "mofiria" system is highly compact and allows a great deal of freedom when using this system in tandem with other products or services, because "mofiria" components can even be positioned laterally without affecting the ability of "mofiria" to do its job.

Figure 1: The reflective dispersion method

Four Processing Algorithms for Finger Vein Patterns

Four algorithms are used together to identify features in blood flow patterns from vein images captured by the system.

Figure 2: Finger vein pattern processing algorithms

--Centerline detection--
The "mofiria" system detects the centerlines of finger veins, which are less likely to be influenced by changes in the weather or physical condition of the individual.

--Straight line pattern authentication--
Differences in finger position can cause rotation or scale distortion. The "mofiria" system enhances authentication accuracy by detecting dominant straight line patterns in the finger veins.

--Low contrast--
Light adjustments are unnecessary since the system can identify patterns even under poor lighting.

--Enhanced noise tolerance--
The system features advanced noise tolerance, ensuring highly-accurate authentication even when using a general-purpose CMOS sensor.

Data Compression Reduces Data Volume to Approximately One-Tenth

Data on vein patterns detected by the system is reduced to approximately one-tenth the original and stored in memory. The resulting small template size enables pattern data to be stored on small mobile devices such as FeliCa cards. Verification data can also be stored on mobile devices.

Figure 3: Vein pattern compression technology

Automatic Finger Position Correction, High Speed Processing

Because Sony's unique algorithms automatically correct for finger movement and rotation, the finger does not need to be held in a specific position. Sony has also succeeded in combining speed and accuracy with user comfort by creating technology that rapidly and accurately detects finger vein patterns from captured vein images. The FRR*1 and FAR*2 ratios are 0.1% and 0.0001% respectively. Processing times are approximately 0.015*3 seconds when using a PC CPU and 0.25*4 seconds when using a mobile telephone CPU.

Figure 4: Automatic finger position correction

*1 False Rejection Rate
*2 False Acceptance Rate
*3 As of February 2009, using an Intel 2.8GHz notebook CPU (based on Sony research)
*4 As of February 2009, using a 150MHz ARM9 mobile telephone CPU (based on Sony research)

Looking Ahead

Sony aims to commercialize the "mofiria" system before the end of fiscal 2009 and is currently studying business options. Sony hopes to incorporate the technology into a variety of mobile devices and gateway security systems, and in developing solutions services.