BIOMETRIC RECOGNITION THROUGH ADAPTIVE FEATURES: A SURVEY

Abstract

—There are a number of biometric identities that are connected with humans, the most prominent 
of which are the fingerprint, palm print, palm vein, finger vein, retina, and iris features.  For the purposes 
of attendance systems, permission systems, and other applications of a similar nature, human beings are 
identified based on their biological identities.  In practically all of the companies that have a medium to 
big number of employees, biometric systems have made their way into the workplace.  The biometric 
systems are also being implemented by a significant number of smaller firms that have a sufficiently larger 
number of employees.  As standalone or hybrid biometric systems, or in conjunction with other 
authentication entities, the biometric systems that are based on the properties of the iris are becoming 
increasingly common.  In order to perform iris recognition, it is necessary to accurately localize the iris 
features from the image of the eye that has been gathered for the purposes of training or testing.  The iris 
extraction process needs the use of two demarcation circles. The first circle is responsible for defining the 
outside boundary, and the second circle is responsible for determining the inner border by detecting the 
outer boundary of the pupil.  Additionally, the angular shift mechanism can be integrated in order to 
investigate the movement of the iris in the image that is provided in order to get precise localization of the 
region of interest that contains the iris characteristic.  A probabilistic classification based on a multi-class 
support vector machine will be utilized by the suggested approach in order to identify the characteristics 
of the iris when contact lenses are present or absent.  It is the goal of the proposed solution to enhance the 
current model in order to achieve more reliable performance. 

,