Using the joint transform correlator as the feature extractor for the nearest neighbor classifier

Financial transactions using credit cards have gained popularity but the growing number of counterfeits and frauds may defeat the purpose of the cards. The search for a superior method to curb the criminal acts has become urgent especially in the brilliant information age, Currently, neural-network-based pattern recognition techniques are employed for security verification. However, it has been a time consuming experience, as some techniques require a long period of training time. Here, a faster and more efficient method is proposed to perform security verification that verifies the fingerprint images using the joint transform correlator as a feature extractor for nearest neighbor classifier. The uniqueness comparison scheme is proposed to improve the accuracy of the system verification. The performance of the system under noise corruption, variable contrast, and rotation of the input image is verified with a computer simulation. (C) 1999 Society of Photo-Optical Instrumentation Engineers. [S0091-3286(99)01401-4].