Font Size: a A A

Research Of Finger Vein Recognition Algorithms Based On Location And Direction Coding

Posted on:2014-11-27Degree:MasterType:Thesis
Country:ChinaCandidate:Q RaoFull Text:PDF
GTID:2268330422460509Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
Traditional identification technology is using markers (key, IC card, etc.)or knowledge (ID, password, etc.) to the identification of humans. Thesetechnologies have disadvantages such as forget and forgery. Biometricidentification is an automatic identification based on the digital information ofhuman’s physiological characteristics or behavior patterns. Biometriccharacteristics are naturally attached on human, nearly invariant with time,and see significant variation between different individuals. All these promisegood application prospect for this technology.Finger vein recognition is a new biometric identification, which ownedhigh stability and uniqueness. As a living feature, finger vein have highanti-counterfeiting capability. Finger vein image are easy to capture andacquisition device have small size. Finger vein recognition has advantagesover other biometric identification and has attracted considerable attentionnow.In this thesis, we firstly establish a finger vein imaging device withnear-infrared light source, by which two databases for finger vein images isestablished. Than we propose a structured personal identification approachusing finger vein Location and Direction Coding(LDC). First, we extract thefinger region by use of edge detection. Second, we make use of the brightnessdifference in the finger vein image to get vein’s locations and directions.Third, we design a two-step threshold segmentation method to extract the veinpattern. Furthermore, finger vein LDC is proposed and performed, whichcreates a structured feature image for each finger vein.Finally, the structured feature image is utilized to conduct the personalidentification on our image database. The identification rate of our methodon the databases is100%and97.3%, and the equal error rate is0.44%and1.53%, which is better than existing method. The experiment result shows that our method has strong robustness and high identification rate.
Keywords/Search Tags:Biometrics, Finger Vein Recognition, Feature Extraction, Threshold Segmentation, Location and Direction Coding
PDF Full Text Request
Related items