Abstract: Security is major concern for today's scenario. Inorder to meet the growing needs of security, several researches are done to provide more privacy to the user. In this paper, an authentication method using image quality assessment is used. There are various biometric methods are available, sclera vein recognition, iris recognition, face recognition etc. Since there are so many techniques are available they are not so reliable. Nowadays biometrics systems are attacked by using fake samples. Since biometrics are concentrated on their accuracy it suffers from the innate disadvantage of time consumption during enrolment and verification process. Even the sclera patterns can be faked to access the biometric systems. Inorder to make the system more efficient, image quality measures are used to detect the input image as real or fake. The proposed method extract 25 image quality measures and confirms whether the input biometric sample is real or fake. This methodology has been compared with QDA and Naive Bayes classifiers.
Keywords: QDA, biometrics, security spoofing, image quality assessment.