jpr.4.3.13885
Evaluation of Hybrid Face and Voice Recognition Systems for Biometric Identification in Areas Requiring High Security
Cheluri Trivikram
Susanna Samarpitha
K. Madhavi
Diana Moses
Journal on Pattern Recognition
2350-112X
4
3
9
16
10.26634/jpr.4.3.13885
Face Recognition, Voice Recognition, PCA, FLD, DTW, SVM, Neural Networks
Biometric identification is a mandatory tool to secure digital information for various industrial, government, commercial, and security applications. Face recognition is a distinct problem and lacks a unique solution applicable to all situations. Face recognition is not effective in identifying individuals in conditions, when a person is using glasses, hats or has a beard etc. Alternative technologies like Iris and retinal scan biometric techniques need sophisticated equipment, which is not financially viable for all applications. Voice recognition methods have low accuracy and are affected by situations where a change in a person's voice due to illness like cold render absolute identification inaccurate. This paper proposes a biometric method implementing multiple techniques i.e., both face and voice recognition technique as an effective identification tool. Identification process using combined biometric methods makes a foolproof security system, thereby leaving no scope for error. A comprehensive assessment of the performance accuracy of several algorithms like Principal Component Analysis (PCA), Fisher Linear Discriminant Analysis (FLD), Dynamic Time Warping (DTW), Support Virtual Machines (SVM), Neural Networks (NN) implemented in face and voice recognition were studied and then corresponding performance accuracy was reviewed. A new technique has been proposed using these performance accuracy results which would help provide the best hybrid method. The proposed hybrid biometric identification method is a viable solution for industries or areas with need for high security of their data systems.
September - November 2017
Copyright © 2017 i-manager publications. All rights reserved.
i-manager Publications
http://www.imanagerpublications.com/Article.aspx?ArticleId=13885