Optical EngineeringPattern identification with an improved synthetic discrimination function
|Format||Member Price||Non-Member Price|
An improved synthetic discriminant function for pattern identification is proposed, which is suitable for discriminating one class from all other classes with distortion invariance. By training the samples and their complementary versions from a specified class simultaneously, the distances between the trained class and all other classes are increased in detection space, and therefore high discrimination results. An application to fingerprint identification is given.