Composite Biometric Authentication Algorithm

Compounding bio-certification is the information fusion of a variety of biological features, such systems integrate the source data provided by multiple biological characteristics to make a more accurate and efficient identification of finds. Technology of Data fusion is a message from the effective integration of all aspects of dealing with new theories and technologies. Data fusion, also known as information fusion, refers to data from multiple sensors to conduct multi-level, multi-faceted, multi-level processing, and thus get more detailed, more meaningful new information. This new information can not be obtained from any single sensor.The basic goal of most of data fusion is to obtain more accurate information through the integration of data than the data obtained from any single, for which explicit fusion algorithm is particularly important.At First, this article analysisese a detailed account of the complex status of biometric research, biometric types and processes and their commonly used fusion algorithm, then construct a human face and iris recognition systems, it introduces the complex process through this system and verify biometric fusion algorithm effectiveness. It introduces the way of multi-template positioning Face through the Face identification, and use the human eye, face pairs of the template locates the face, collect facial feature by combining with Adaboost algorithm and principal component analysis. It Introduces the iris location through the iris recognition from the iris to the iris image enhancement, normalized as the number of iris feature extraction process, uses randomized Hough transform based fast detection way of circle to detect the internal and external circular position of the iris, extracts the iris phase-coded information through the complex two-dimensional Gabor transform.On the extracted features of face and iris information, we put forward the data fusion algorithm by using the Combination Fisher discriminated and majority of voting.At first, use the Fisher discriminated function to combine two kinds of feature information fusion, and finally determine the final by majority voting method,. At last, test and verify the feasibility of the algorithm by experiment and then get a higher recognition rate.