Development of a Enhanced Ear Recognition System for Personal Identification

This paper examines the development of a Match Region Localization (MRL) Ear Recognition System (ERS). Captured ear images were pre-processed through cropping, and enhancement. The pre-processed ear images were segmented by MRL segmentation algorithm, where the pre-processed images were sub divided into 96 sub images. Principal features from the segmented ear were extracted and used for template generation. K-nearest Neighbour classifiers with Euclidean distance metrics were used for the classification. The developed ERS demonstrated a recognition accuracy of 96.00%. The developed system can be tested on other available public ear databases to allow for cross-database comparison. Moreover, the system can still be improved for further reduction of ERS.