Identifying a person using biometrics is increasingly becoming popular these days. While using biometrics, a number of modalities are available in which face recognition, fingerprint and palmprint identification are widely used. Among them, palmprint is the widely accepted solution because of its richness of features and the low cost involved in acquisition. In this paper, a novel biometric authentication scheme that uses Contourlet transform for performing analysis has been proposed. This transform is a pyramidal directional filter bank which can be applied to capture both local and global features of palmprint. In this process, a set of features are extracted first. Now, dimensionality reduction is carried out by calculating the energies for each subband and then uses them to minimize the vast amount of coefficients generated. Finally, palmprint matching is performed by comparing the input palmprints with the available palmprints stored in the database. The available palmprints are classified using an Enhanced Multi-class Support Vector Machine (EMSVM) algorithm so that in the testing phase the input palmprint can be compared with a suitable palmprint. This algorithm has been tested using the CASIA database of images which were taken using self-developed palmprint recognition device. © EuroJournals Publishing, Inc. 2011.