Multimodal Biometric Identification

With the development of human society,identity authentication and security issues has become more and more important.Human biological characteristics include human face,iris,fingerprints,veins,sounds,etc.These biological characteristics are unique to each individual human beings and are different from other individuals,the study of these characteristics can be beneficial to human identity identification.Gabor transform is a very good feature extraction algorithm.Combining Gabor transform and wavelet transform can effectively improve the recognition accuracy,but both of them have the problem which is long recognition time.When algorithm of DCT is introduced into the recognition process,the recognition time can be effectively reduced,while the recognition rate is still very impressive.The idea of Gabor transform is very similar to the human retina recognizing things,so it has become the main reason for people to study it.Gabor wavelet transform is a class of wavelet transform,but inherits the advantages of Gabor transform,the idea is to convolute the entire image,and takes into account each pixel of the image,and also consider the correlation between pixels,which makes The recognition rate of such algorithms is quite impressive.However,precisely because Gabor transform considers all the pixels,and most of the pixels are not feature points,which is equivalent to do a lot of useless work,making recognition time longer.The advantage of DCT algorithm is to effectively compress the image,and can remove the correlation between the pictures,which can make up for the shortcomings of Gabor transform.First use DCT algorithm for image processing,which makes the main energy of the picture to concentrate in the upper left corner of the part,in fact,it can make the feature point and non-feature points effectively distinguish,and then intercept the upper left corner of Gabor wavelet transform,the process actually is to capture the main feature points of the image,remove most of the non-feature points,which effectively saves the workload,making the recognition time significantly reduced.Face recognition mainly considers the influence of light,attitude and occlusion(such as wearing glasses),etc.The improved Gabor wavelet transform has some advantages in solving these influences.Iris recognition pre-processing is a difficult task,including the iris localization and normalization,that is,iris extraction is a key part of iris recognition.Fingerprint recognition pretreatment is also a difficult task,the pretreatment process includes fingerprint enhancement and refinement,thisprocess is also a key part of fingerprint identification.In a word,through the research on face recognition,iris recognition and fingerprint recognition,comparing the improved Gabor wavelet transform and the improved one,it can be found that the improved algorithm can effectively reduce the recognition time and maintain a good recognition rate.