Performance Evolution Ear Biometrics Based on Features from Accelerated Segment Test


  • B.A. Mohammed and Z.M. Abood


Ear Biometric, Detection technique, FAST  Descriptor, Bilateral Filter, Histogram Equalization.


In the field of image processing, feature extraction is very important. Various image pre-processing procedures, including scaling, downscaling, resizing, normalization, etc., are applied to the sampled image before features are acquired. Features that would be relevant for image classification and recognition are then extracted using feature extraction methods. Many problems arose in biometric methods (fingerprint, iris, face), which led to the search for new biometrics to identify a person, avoid disease obstacles, continuous change with age, and others. The aim of this work is to present an ear system based on the power extraction of biometric field features of the ear. The proposed system will extract the robust features using the FAST method. This system reduces sprint time while maintaining its accuracy. This is done in two stages, the first stage, is the application of several pre-processing techniques which included the process of a) conversion to grayscale, b) binary filter, c) image brightness, histogram equalization, and image noise reduction. In the second stage, the method of transforming the fixed attribute variable FAST is used to determine the extraction of the force field feature. Where 70 % images were used for the purpose of training and 30 %for the purpose of testing. The accuracy of these figures was tested, reaching 95%, and therefore they can be adopted in the classification process in the future by choosing any algorithm.