Efficient algorithms for face recognition based on spatial and transform domain techniques

Abstract

The biometrics are used to identify human being for real time newlineapplications. The biometric identification is efficient compared to conventional methods of identification as biometrics are body parts of human beings and behavior of a person. The biometric recognition systems are designed using spatial domain techniques, transform newlinedomain techniques and hybrid domain techniques. The designed biometric systems are implemented using Reversible Logic Multipliers in real time. Chapter one presents the introduction to Biometrics. The block diagram of biometric system is explained. The feature extraction based on spatial domain and transform domain is discussed. The newlinedefinitions of performance parameters, applications of biometrics, motivation and objectives are discussed. The literature survey on Face recognition based on different existing techniques are discussed in chapter two. The face recognition based on spatial domain and transform domain feature extraction techniques are described. The various existing matching techniques are also discussed. The hybrid domain based face recognition using DWT, FFT, CLBP and histogram is proposed in chapter 3. The face databases newlinesuch as JAFFEE, ORL, Indian male and Indian female are used to test the performance of the system. The DWT is applied on face images and considered only LL sub band and is resized to the original face image. The Laplacian filter is used on the resized LL band to obtain LL coefficients of face images. The face image pixel intensity values are newline