18-06-2012, 12:03 AM
Facial recognition technology is more efficient
in the field of biometrics. Fingerprints are at
risk of counterfeiting, as well as to reach
wrongly prevent access if the fingers are dirty
or wet. The iris recognition devices are too
invasive. Can generate serious diseases or
conditions. The voice recognition devices can
not be successful, especially when you have a
sore throat. Maybe facial recognition is not
perfect, but is more advanced than most
existing identification methods.
Lets talk about facial recognition systems. In
the facial recognition systems 2D images are
most commonly used rather than the 3D
images. though both have advantages and
disadvantages. For some reason, the 2D work
better under lighting conditions, but often
affected by changes in gestures or expressions.
3D recognition systems are not as common, but
tends to become increasingly cheap and fast. On
the downside, the database of 3D images are
very small compared to 2D.
The fundamental difference between both
models is the most obvious: the system
considers only two dimensions 2D face, while
the generated 3D images that resemble the real
thing. Tend to think that the 3D system in itself
guarantee better results, but this has not been
scientifically proven. From the point of view of
experts, the 3D system is too expensive for
most people, although it is true that it tends to
become cheaper. On the other hand, sometimes
its effectiveness depends on the disparity of
light.
Experiments have shown that better results are
achieved with the combination of both
technologies. The use of many different
biometric variables ensures higher accuracy
than the best sensor for one of them. In a recent
study found that combining 2D and 3D can
achieve a hit rate of 98.8%.
In conclusion, the 3D tends to be popular and
promises to be gaining new benefits over 2D,
but the latter is the most widely used model
because it has many advantages, has a much
greater development time and is constantly
improving. The 3D can provide better results in
certain cases, but still requires much research
and development for its consolidation. Both
systems can coexist and complement smoothly.