International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395 -0056
Volume: 04 Issue: 03 | Mar -2017
p-ISSN: 2395-0072
www.irjet.net
A REVIEW PAPER ON PERSONAL IDENTIFICATION WITH AN EFFICIENT METHOD OF COMBINATION OF LEFT AND RIGHT PALMPRINT IMAGES Dipalee M. Kate1, Nita Lagdewar2 1Assistant
Professor, Dipalee M. Kate, Dept. of ECE, Priyadarshini Bhagwati college of Engineering, Maharashtra, India
2Nita
Lagdewar, Dept. of ECE, Priyadarshini Bhagwati college of Engineering, Nagpur
---------------------------------------------------------------------***--------------------------------------------------------------------Fusion in multimodal biometric system can be performed at ABSTRACT- Multibiometrics provides high standard four levels: In the Image Level, different sensors are usually security and better accuracy than any other single required to capture the image of same biometric. At Decision biometric. Palmprint identification achieves a high Level, fusion at decision level is too rigid. It only abstracts accuracy because it contains principle curves, wrinkles identity labels decided by different matchers. At Feature and texture and also rich texture and miniscule points. In level, it involves the use of feature set by concatenating this paper, we perform multibiometric by combining left several feature vectors to form large 1D vector. It provides palm print and right palmprint at matching score level better identification accuracy than fusion at other levels. fusion.The first two kinds of score were obtained from palmprint identification method. For third kind of score, At the Matching score level, the final matching score is we propose a special algorithm which takes the nature of generated from three kinds of matching scores. The first and left and right palmprint images. It can properly exploit second matching scores obtained from left and right palm the similarity between left and right palmprint of same print reps. The third kind of score is calculated based on the object. crossing matching between the left and right palm print. In this paper, we proposed technique which combines the left with right palmprint at the matching score level. The framework contains three types of matching scores which are obtained by the left palmprint matching, right palmprint matching and crossing matching between the left query and right training palmprint and they are fused to make the final decision. It combines the left and right palm print images for identification and also properly exploits the similarity between the left and right palmprint of the same subject. The proposed framework can integrate most conventional palm print identification methods for performing identification and can achieve higher accuracy than conventional methods.
Keywords: Multibiometrics, SIFT, Multimodal.
1 .INTRODUCTION During the last years there has been an increasing rise of automatic personal recognition systems. Palmprint based biometric approaches have been intensively developed. Because they possess several advantages over other systems. Palmprint a large inner surface on our hand contains many line features. For example: principal lines, wrinkles and ridges. Because of the large surface and the rich line features, we expect palm prints to be robust to noise and to have high in individuality .To overcome the limitation of unimodal biometric technique and to improve the performance of the biometric system, multimodal biometrics. Use of multiple biometric indicators for identifying individuals Known as Multimodal biometrics. Using an effective fusion scheme can significantly improve the overall accuracy of the biometrics system obtained from the combining evidence from different modalities.
© 2017, IRJET
|
Impact Factor value: 5.181
2. LITERATURE SURVEY 1. R.K.ROWE, “A multispectral whole-hand biometric authentication system”, Refer to [7], in 2007, A multispectral whole-hand biometric system has been developed.Its main objective is to collect palm print information with clear fingerprint features and preprocessing .The speed of feature extraction is very low.And
|
ISO 9001:2008 Certified Journal
|
Page 863