Finger Knukle Print Recognition based on the Combination of the Multi Shift Local Binary Pattern Descriptor with Discrete Fourier Transform

The old security systems used have proven to be not very efficient as they are easy for hackers to obtain. The most crucial component of computer systems is user safety. The use of biometrics is more secure in identifying identity. There has been more interest in Finger Knuckle Print (FKP) as a feature of biometrics. The reason was due to the high efficiency and accuracy that this measurable feature implied. This paper presents a blending approach that combines information from two algorithms, Multi shift Local binary patterns (MSLBP) and Discrete Fourier Transform (DFT) method, to create a new vector of FKP. In the classification process, Support Vector Machine (SVM) was used. The PolyU FKP database was used in the experiments. The outcomes of the evaluation shows that the proposed approach improves the performance of FKP compared with a number of existing approaches.