PERSONAL AUTHENTICATION BASED ON HAND DORSA VEIN IMAGES

Shah Hida, Shah Jahan (2010) PERSONAL AUTHENTICATION BASED ON HAND DORSA VEIN IMAGES. [Final Year Project Report / IMRAD] (Unpublished)

	PDF Shah Hida (dsva).pdf Restricted to Repository staff only Download (816kB)
	PDF Shah Hida 24 pgs.pdf Download (9MB)
	PDF (Please get the password by email to repository@unimas.my , or call ext: 3973 / 3942 / 3933) Shah Hida ft.pdf Restricted to Registered users only Download (45MB)

Abstract

Personal authentication is a way to verify or identify a person from a given data. There are many way are use to verify a person. For example based on hand geometry, fingerprint, iris, face and hand vein. While in this project, hand dorsa vein images are used to verify the person. Every person has a unique set of blood veins in his or her hand. Veins are very complicated therefore contain lots of different lcatures that help to identify a person. In this project there are 2 categories that is image preprocessing and post-processing. Image preprocessing is mainly to make the images clearer while post processing is a process to verify the images. In image preprocessing, there are 3 steps to process it that is noise reduction, image segmentation and image extraction to get the region of interest. As for post processing backpropagation neural net are used as a method to train and testing the image. Then Principle Component Analysis is use to extract the feature vector. Only one feature vector was selected randomly for each person. Then a image of each person will be trained one by one. The value of the weight taken from the training will he updated in a database. 60 images for training which consists of 45 images from the same person that have in database while 15 images are other person images. Images verification was done by one- to- many. The performances were evaluated using False Acceptance Rate (FAR) and False Rejection Rate (FRR)

Actions (For repository members only: login required)

View Item