As an emerging biometric device that is expanding rapidly, iris recognition is a new identification technology that is founded on the authentication of an individual through the recognition of the iris pattern, unique to every individual (1). Unique features of the iris is due to its high complexity, combined with high degree of randomness and variation, giving each individual a distinct pattern for recognition (3).
Contrarily to facial recognition that is much less accurate due to the continuous facial changes occurring as one ages, the iris of a person is more likely to be constant because most of its structural characteristics are formed before the end of one's gestational period (2) and although it is exposed to the external environment, this structure is also protected by eyelids to prevent damages (2).
This biometric technology works by taking a monochromatic picture of an individual's iris, taking into account the position of the iris to leave out eye lashes or other noises that would decrease the recognition accuracy and efficiency (1).
The pattern captured by the camera is then translated into an IrisCode (1). Recognition works by comparing an individual's IrisCode with those previously recorded into the database and the degree of difference is then measured as Hamming Distance (1). This statistical independence test implicates that a certain threshold (difference of IrisCodes below 1/3 of the bytes) needs to be reached in order to confirm that the pair of IrisCodes originated from the same iris, and thus same individual (1).
(1) http://www.biometrics.gov/Documents/IrisRec.pdf
(2) http://www.cl.cam.ac.uk/~jgd1000/irisrecog.pdf
(3) http://www.icdri.org/biometrics/iris_biometrics.htm
Contrarily to facial recognition that is much less accurate due to the continuous facial changes occurring as one ages, the iris of a person is more likely to be constant because most of its structural characteristics are formed before the end of one's gestational period (2) and although it is exposed to the external environment, this structure is also protected by eyelids to prevent damages (2).
This biometric technology works by taking a monochromatic picture of an individual's iris, taking into account the position of the iris to leave out eye lashes or other noises that would decrease the recognition accuracy and efficiency (1).
The pattern captured by the camera is then translated into an IrisCode (1). Recognition works by comparing an individual's IrisCode with those previously recorded into the database and the degree of difference is then measured as Hamming Distance (1). This statistical independence test implicates that a certain threshold (difference of IrisCodes below 1/3 of the bytes) needs to be reached in order to confirm that the pair of IrisCodes originated from the same iris, and thus same individual (1).
(1) http://www.biometrics.gov/Documents/IrisRec.pdf
(2) http://www.cl.cam.ac.uk/~jgd1000/irisrecog.pdf
(3) http://www.icdri.org/biometrics/iris_biometrics.htm
No comments:
Post a Comment