GENERAL METHODOLOGY FOR SIMULTANEOUS REPRESENTATION AND DISCRIMINATION OF MULTIPLE OBJECT CLASSES

We address a new general method for linear and nonlinear feature extra ction for simultaneous representation and classification. We call this approach the maximum representation and discrimination feature (MRDF) method. We develop a novel nonlinear eigenfeature (NLEF) extraction t echnique to represent data with closed-form solutions and use it to de rive a nonlinear MRDF algorithm. Results of the MRDF method on synthet ic databases are shown and compared with results from standard Fukunag a-Koontz transform and Fisher discriminant function methods. The metho d is also applied to an automated product inspection problem (discrimi nation) and for classification and pose estimation of two similar obje cts under 3-D aspect angle variations (representation and discriminati on). (C) 1998 Society of Photo-Optical Instrumentation Engineers. [S00 91-3286(98)01103-9].