Multimodal Visual Biological Feature Embedded Recognition System Design And Implementation

Due to the large population and its mobility, Biometrics has become one of the mostimportant authentification techniques nowadays. As reflected by the investigation and theresearch, multimodal biometrics techniques have shown more stability and security thanmonomodal biometrics as they take more than one physiological or behavioral inherentcharacteristics of an individual into account to identify. With the development of theembedded technology we claim more portable and practical system. Embedded technologygenerally have the features of exclusivity, temporality, flexibility etc. Therefore, it couldadopt all the working environment such as multimodal biometrics system. In this paper itgeneralizes a new multimodal biometrics system based on embedded technology.The essential fuction of our multimodal biometrics system based on Embedded iscapturing visual biological characteristics (fingerprint, finge vein and so on.), and thenconverting all the data to image. So we could use these data for further establishtion of thehardware and software of the system.According to the requirement of the multimodal biometrics based on Embedded, weuse the processor TMS320DM6446from TI company and implemented in Linux operatingsystem. We also complete the hardware and software of the system. And the latter dependson the former.This paper contains the hardware platform construction, Linux operation system basedon embedded technology, and video/image capture etc. We just focus on the correspondinginterface of Linux driver program. We just require the system to achieve multimodalbiometrics capture function, therefore, we use dual video/image acquisition interface forvideo/image capture module design. So this paper focuses on its description, including thehardware circuit design and software drive design. In the process of the video/image captureinterface design, because of the characteristics of the processor TMS320DM6446structureand requirement of the video/image capture. The research is concerned about video/imagedriver program of the V4L2structure. Eventually we complete these function asshown.V4L2drive, VPFE drive and a complete video capture interface drive or digitalimage capture interface drive. All these function construct a complete video/image capturedriver program.The system accomplishes the hardware and software platform if multimodal biometricsidentification device based on the embedded technology and biological feature capturefunction. They provide effective technical support for the multimodal biometrics algorithm.