Research On Algorithm Of Establishing Hand - Palm Network Database

With the development of biotechnology and computer technology, thebioinformatics become more and more popular in the field of computer research. Alot of researches, such as finger-prints, palm prints, face, ears and other peoplebioinformatics have become focus in the computer bioinformatics research, and thedomestic and foreign experts pay attention to the palm biometrics. Today palmbiometric has a wide application, such as palm-print identification and the Chinesemedicine theoretical digitized research. The palm-print network database is the basicfor palm-print biometric applications; these applications are inseparable from thepalm-print database. The complete standard network palm-print image databaseapplication system will speed up the palm research and application. Today someindividual research institutions have established palm-print database for recognitionresearch, but there is rarely database for Chinese palm-print diagnostic research. Asthe birthplace of traditional Chinese medicine theory, build a palm database forChinese medicine will promote the research and application on traditional Chinesemedicine digitized, it has significance social and scientific value.This paper is to establish the palm database for Chinese traditional diagnosisresearch and related applications study, mainly included image capture and imagestandardization, the image acquisition is the focus of this article research. Networkusers have to use ordinary camera or CMOS camera to capture color palm images andcapture true palm color; Image standardization include as follow: Image de-noising,size reset, rotation correction and palm feature region extraction algorithms. Specificstudies and the results obtained are as follows:(1) Study the cameras white balance algorithm under the natural light officeenvironment with different aperture sizes, different exposure time, manually andautomatically set. Mainly studied several classic white balance algorithm palmcolor consistency.(2) Use natural office environment light with ColorChecker characterization method to revise the color of palm image, we study how to capture consistencycolor use different camera with different settings for palm. The experimentalresults show that the characterization camera method is very effect andconsistency, which under different aperture, exposure settings as well as differentcameras. After the correction, the chromatism of palm color is reduce from thepre-correction square color6.22CIELAB color difference units to0.84CIELABunit.(3) In order to extract palm-print characteristic, we have studied imagestandardization preprocessing: include image de-noising, size reset, rotationcorrection. At the same time, we have study the palm’s ROI extract algorithms,we summery those algorithms mainly into two types: based on the maximuminscribed circle and based on feature points extraction algorithm. We didexperiments use several palm image database to verify this two main algorithms.