Research On Fusion Recognition Method Of Palm Vein And Palmprint Based On Single Infrared Palm Image

In recent years,the domestic and foreign scholars have done a lot of research about palm for fusion recognition technology,it mainly includes palmprint and hand shape for fusion recognition;palmprint,hand shape and fingerprint for fusion recognition;hand dorsal vein and finger joint for fusion recognition;finger vein and finger dorsal for fusion recognition;fingerprint and finger vein for fusion recognition;palm vein and palmprint for fusion recognition,and so on.The palm vein and palmprint for fusion recognition has become research direction in recent years.However,there is little research that use single image to realize palm veins and palmprint for fusion recognition.In this thesis,the near infrared palm image database of Hong Kong polytechnic university is used to study the method of multi-feature for fusion recognition.The research contents are as following:extracting the palm vein structure image and the palmprint structure image form the single near infrared palm image;proposing the unimodal recognition method of palm vein and palmprint in the near infrared palm image;using the palm vein and the palmprint for fusion recognition based on fixed weight;using the palm vein and palmprint for adaptiveweight of fusion recognition based on image quality;using the palm vein and the palmprint for adaptive weight of fusion recognition based on image similarity coefficient.This paper is structured as follows:The method of palm vein structure image and palmprint structure image were extracted and enhanced separately from the single near-infrared palm image is proposed.The palm vein and palmprint structure information actually exists in the same near infrared palm image.Due to the palm vein and palmprint structures stay in the different ranges of pixel values,and their local texture information is also different,the palm vein and the palmprint can be separated from a single near infrared palm image and then enhanced individually.An improved self-guided filter algorithm is proposed to remove the palmprint structure,and the anti-fuzzy detail enhancement model is designed to enhance the image of the palm vein structure.An improved block enhancement algorithm is proposed to enhance the palmprint structure and remove the palm vein structure at the same time.Then,the Sobel operator unsharp masking algorithm is used to highlight the main line structure of the palmprint.The method of unimodal recognition of palm vein and palmprint in the near infrared palm image is researched.The features of palm vein and palmprint are extracted respectively by wavelet decomposition,and using similarity coefficient between two samples for classification.The palm vein recognition rate of the near infrared palm image is 99.23%when the improvement of the self-guided filter parameters(r1,r2)were(2,16)and the detail enhancement parameter t was 5.The palmprint recognition rate of the near infrared palm image was 94.00%when the block size R was 9 and the detail enhancement parameter ? was 50.The method of palm vein and palmprint for fusion recognition based on fixed weight is researched.The recognition thresholds of the palm veins and palmprint are calculated respectively.The palm vein and palmprint for fusion recognition based on recognition thresholds are weighting.Under this circumstance,the fusion recognition rate of the palm vein and palmprint was 99.67%when the threshold weight value of palm vein was 0.6 and the palmprint was 0.4.The method of the palm vein and palmprint for fusion recognition based onimage quality is proposed.The palm vein structure image and the palmprint structure image are evaluated from the edge point/fuzzy point,high frequency information/low frequency information,and the average contrast of image.According to above three factors,the method of the palm vein and palmprint for fusion recognition based on image quality is proposed and the fusion recognition rate has reached 99.69%.The method of palm vein and palmprint for fusion recognition based on image similarity coefficient is proposed.The higher the similarity coefficient between two images is,the higher the confidence level of the two images belongs to the same class is,and the research shows that the average similarity coefficient between the palm vein structure images is higher than the palmprint images.The fusion recognition of palm vein and palmprint based on image similarity coefficient is proposed,and the fusion recognition rate has reached 99.71%.