Increasing accuracy and reducing time of face recognition with Euclid norm

Abstract

Biometric parameters are used largely than smart cards, passwords or others in authentication processes. They differ from other methods with stability and indispensable features. In practice, fingerprints, face, iris, and gate based biometric authentication methods are used. Researchers compare biometric parameters by reliability to forgery, no changing, no repeating and convenient to use. Authentication based on face is very popular and suitable to parameters. The major advantage is the only physiological biometric that can be reliable mark at distance and, so, the verification of the users can happen without their accurate interaction with the sensor or their knowledge. Face recognition solutions are used for controlling attendance in airports, stadiums, railway stations, education and other organizations. The main problems of face recognition algorithms are recognition time and accuracy. We will solve this problem by changing Frobenius norm to Euclid norm in proposed method. Ways of optimization and several recommendations for collection face database are proposed.