We propose an individual recognition system using eigenear images. Although face images change owing to changes of emotion, growth, expressions, infirmity, and so on, ear images cannot be affected by them. Our system utilizes principal component analysis for dimensionality reduction, and Mahalanobis' Distance for individual recognition.
Experimental results show that our system can recognize individuals with the high recognition rate 99.7% - approximately the same rate of an existing method based on Genetic Aleorithm(GA) - in shorter time than that of the GA-based method.
View full abstract