An Efficient Human Ear Recognition Scheme Based on Robust Features Extraction and PNN

Abstract

Nowadays, the technology of biometrics has been gained more popularity for the security and other purposes. The human ear is a new and popular biometric technology which holds truly considerable merits over other biometrics such as face, fingerprints, retinal, and iris scans. The schemes of human ear recognition represent a significant research field to identify the persons since the ear biometric includes stable and rich features and the ear structure is unaffected to change with the facial expressions and age.In this paper, an efficient and robust scheme has been proposed for human ear recognition based on several features extraction (Features from Accelerated Segment Test (FAST), and histograms of oriented gradients (HOG)). A Probabilistic Neural Network (PNN) is utilized for recognizing the user via matching the user's extracted features with the database of features.The proposed scheme is evaluated by utilizing AMI ear database, and when the FAST feature extraction are used, the better rate of recognition is achieved (98%). The obtained results demonstrate that the proposed ear recognition scheme provides better performance than other recently existing ear recognition schemes.

Keywords: Human Ear Recognition, Histograms of Oriented Gradients (HOG), Features from Accelerated Segment Test (FAST), Probabilistic Neural Network (PNN).