Algorithm can turn 2D images into 3D biometrics
Researchers at Florida Atlantic University have developed a way to construct 3-dimensional faces from flat 2-dimensional images; the computer algorithm analyzes an image and constructs a 3-dimensional model based on viewing angle and illumination; using this technique computers can more accurately identify individuals as computers often have a hard time distinguishing between individuals with flat images; this algorithm can also be used to identify criminals caught on security camera
Researchers at Florida Atlantic University have developed a way to construct 3-dimensional faces from flat 2-dimensional images.
Xin Guan and Hanqi Zhuan published their results in the January issue of the International Journal of Biometrics.
The two scientists developed a computer algorithm that has the capability to analyze an image and generate a 3-dimensional face using the viewing angle and illumination.
The researchers say that the value of flat images is limited as human faces often share too many similar characteristics making it difficult for computers to distinguish individuals.
The introduction of these 3-dimensional images could greatly boost accuracy of identification in large databases and can be used in conjunction or even replace fingerprint, iris, voice, or DNA biometric identification methods.
Researchers also say law enforcement officials could use the same algorithm to identify individuals caught on security cameras or even missing persons.
