Ear Recognition for Human Identification

Abstract

Ear biometrics is a developing method for human identification due steadiness of human ear features throughout the lifetime. This review paper discusses about some of the previous works done on the field of human recognition using ear. In one of the methods the ear segmentation is done using fourier descriptors and the features are extracted using Gabor filter and classification is done using KNN classifier. In the other method snake model is applied on each of the ear image to generate a mask which is ear specific. Then this mask is applied on the corresponding gray scale ear image to crop the ear region then geometrical features are extracted such as eucledian distance, centroid, mean median, co-ordinates of centroid. For this method also KNN classifier in used for classification. In the proposed method the idea is to use snake model to generate a mask for ear segmentation. The database to be used is the IIT Delhi ear database which is publicly available. In this paper the methods are explained based on the review papers. Then the comparision is to be done between various methods.