Vous êtes sur la page 1sur 1

Development of a palm-vein illumination and

sensor system

Stephen D. Wolthusen and Christoph Busch

Palm-vein detection is a biometric system which has mainly been devel-


oped in Japan and is already in commercial use, e.g. in nance and health
care. The pattern of veins in a human palm is a randotypical feature that
emerges during tissue growth and is therefore likely to be distinct even for
identical twins. At the same time, it oers a relatively large feature vector,
resulting in attractive error rates being possible.
In this proposed project, a palm vein pattern sensor is to be developed,
which will require a combination of physical prototype construction and soft-
ware engineering.
Palm veins do not exhibit high contrast in visible light, but are detectable
in the near-infrared spectrum as the blood and blood ow induces tempera-
ture dierences with the surrounding tissue. The contrast can be enhanced
by further near-infrared illumination.
As part of the project, a camera and illumination mechanism is to be
developed from standard or laboratory components, resulting in a series of
palm images being recorded, and the sensor system is to be optimized for
enhancing the contrast of the palm vein image using changes to the sensor
system.
The resulting raw sensor data is then to be enhanced using standard (au-
tomated) image processing techniques, comparing the result to commercially
available sensor data where possible.
An area to be studied further is the quality of the data acquired in
this way that climate conditions such as humidity and temperature aect
such sensors, e.g. in cold weather, identifying potential limitations regarding
application areas.
An optional extension may then identify ways in which liveness tests
can be integrated into this process, particularly using motion video capture
either to identify blood circulation or involuntary and prompted motions of
the hand.

Vous aimerez peut-être aussi