Research On Hand Shape Recognition System Based On DSP

Recently, with the development of economy, society, technology, and theimprovement of people’s standard of living, the people’s safety awareness hasimproved. Because the biometric identification technology has higher recognition rate,it is displaying the traditional verification technology, the hand shape recognition isone of the most important biometric identification technology. Hand shape recognitiontechnology has the advantages of facilitating the hand Image collection, the highdegree of acceptance, simple algorithm and fast recognition speed, so more and morepeople pay attention to transgenic silencing.The biometric identification products also play a significant role in more andmore public places, because of the bigger size of the biometric identification productsbased on PC, the applications of these products are limited, and the embeddedbiometric identification systems are welcomed by the market because of its simplestructure and easy operation. According to the development trend of embeddedsystems, the embedded hand shape recognition system has a great market potential. TIDM642processor is the high-performance chip which is specifically aimed at digitalimage processing, the paper designs the hand shape recognition system based onDM642, and the main research contents are as follows:(1)For the problems of the finger root deformation caused by hand locationinaccurate and a big error of Hand shape feature extraction, Taking advantage of itsrole as Finger contour high stability, a new algorithm of hand location is designed. Analgorithm is presented for Hand shape recognition foreign to the finger root contour.This method first locates the finger central axis, then Extracting finger geometricfeatures that are none of the finger root contour. The final statistical characteristicdifferences between different fingers to recognize hand shape.(2)The paper designs the hand shape recognition system based on Matlab. Thesystem has follow functions: image collection, image storage, image processing,feature extraction and feature matching.(3)The paper introduces the hardware platform based on DSP, including systemselection, system structure, data storage modules, image input and output modules.(4)The system framework and software design process are designed, the videocapture program based on DSP/BIOS are developed, successfully transplanting thealgorithm of hand shape recognition to the DSP system. (5)This paper introduces in detail the system debugging schemes and analysesthe performance of the hand shape recognition system.