Principle Component Analysis (PCA) is one of the most important statistical techniques for feature extraction. It is a way of identifying characteristics feature in iris, and expressing the data in such a way as to highlighting their similarities and differences 13. PCA is also known as Eigen analysis. This technique extracts the main variations in the feature vector and allows an accurate reconstruction of the data produced from the extracted feature values. PCA identifies the strength of variations along different directions in the iris which involves computation of Eigen vectors and corresponding Eigen values. The Eigen vectors with largest associated Eigen values are the principal components and correspond to maximum variation in the iris.
K-Nearest Neighbor is the simplest of all classifiers that classifies objects based on the nearest training neighbors in the feature space. It uses Euclidean distance as a distance metric. This classifier can be used to compare two templates, especially if the template is composed of integer values. Object is classified or recognized based on the majority number of its neighbors, and is assigned to the class that is most common among its k nearest neighbors. If 1-nearest neighbor is considered, then the object is assigned to the class of its first nearest neighbor; generally larger values of k are considered to reduce the effect of noise on the classification. In terms of Euclidean distance, the difference, d between the enrolled iris templates and matching iris template k, is given as 14:
The template generated in the feature extraction stage needs a matching metric to measure the similarity between two iris templates. This metric gives one range of values when templates generated from the same eye are compared and another range of values when templates generated from different person’s eye are compared; so that we can decide as to whether the two templates belong to the same or different persons.
Iris recognition system is a rising field of information technology that uses human iris to identify them. By calculating the iris feature, it is possible to identify each individual in a population. The reason why iris recognition is an attractive field is due to the fact that iris feature cannot be forgotten or lost, they are difficult to copy, share and distribute and they require the person to be present at the time of authentication.
Experimental results have demonstrated that the proposed method achieves good performance in run time. This confirms that the proposed combination strategy of feature extraction is suitable for reducing dimensionality of iris templates as well as run time for matching.