Human Ear Identification using Vector Quantization Algorithms

Abstract: Biometrics refers to the identification of humans by their characteristics or traits. An important aspect of the characteristic is that it should be unique for every individual to enable identification by a biometric system. In this research paper, human ear has been used as a suitable characteristic for identification purposes. A raw image is taken an input and an edge detection operator has been used for enhancement to obtain the region of interest. Several Vector Quantization algorithms such as Linde-Buzo-Gray (LBG), Kekre‟s Error Vector Rotation (KEVR), Kekre‟s Median Codebook Generation (KMCG) and Kekre‟s Fast Codebook Generation (KFCG) have been applied to extract the unique features. The same process has been applied to another database containing images of the same ears for testing purposes. For each algorithm, accuracy has been calculated based on the number of correct identifications. The performance, accuracy and complexity of the algorithms are compared.

Keywords: Image Processing, Biometrics, Vector Quantization, Image Pre-Processing, Edge Detection, Linde-Buzo-Gray (LBG), Kekre‟s Error Vector rotation (KEVR), Kekre‟s Median Codebook Generation (KMCG), Kekre‟s Fast Codebook Generation (KFCG), Complexity.