Thermal imaging of face vein patterns enters the biometric discussion
The development of new modalities are always exciting for the biometrics community, and an Indian research team has just come out with a new method for identification using thermal imaging and the blood vessels just below the surface of your face.
Reported in ScienceDaily, a team at Jadavpur University in Kolkata, India, has written about the new development in an upcoming issue of the International Journal of Computational Intelligence Studies, and explains how the vein pattern in the human face works as a unique identifier.
Based on an infrared scan, the new system analyzes minutiae of blood vessels and boasts an accuracy rate greater than 97 percent.
Vein pattern analysis is nothing new for biometric identification. Many fingerprint and palm scanners currently analyze sub-dermal vein patterns for identification, and also in these cases, contact isn’t necessary with the sensor as infrared is used.
Adoption has yet to be determined, but it should be said that many unique identifiers have been discovered and very little is being done with them in terms of development or adoption. This is usually for a few reasons: expensive equipment, cumbersome setup, invasive data capture or low accuracy rates.
For example, as we recently reported, a computer scientist from the Lawrence Technological University in Michigan has developed a system for biometric identification using people’s knees, though it requires an MRI system, which is a significant setup and could explain why it hasn’t caught on as a common biometric modality. Also, so far, tests have only shown a 93 percent accuracy rate, which is low.
Though it has yet to be seen, something that may inhibit the adoption of this new thermal imaging face vein system is that an infrared camera is required.
Facial recognition and voice biometrics are growing at incredible rates in the global market and adoption is really picking up. That’s thanks to a number of contributing factors, but it’s important to note that both voice biometrics and facial recognition can be performed with basic tools.