Multilinear principal component analysis for iris biometric system

<p>Iris biometric modality possesses inherent characteristics which make the iris recognition system highly reliable and noninvasive. Nowadays, research in this area is challenging compact template size and fast verification algorithms. Special efforts have been employed to minimize the size of the extracted features without degrading the performance of the iris recognition system. In response, we propose an improved feature fusion approach based on multilinear subspace learning to analyze Iris recognition. This approach consists of four stages. In the first stage, the eye image is segmented to extract the iris region. In the second step, wavelet packet decomposition is conducted to extract features of the iris image, since good time and frequency resolutions can be provided simultaneously by the wavelet packet decomposition. In the next step, all decomposed nodes or packets are arranged as a 3^rd order tensor rather than a long vector, in which feature fusion is directly implemented with multilinear principal component analysis (MPCA). This approach provides a more compact or useful low-dimensional representation directly from the original tensorial representation. Finally, a discriminative tensor feature selection mechanism and classification strategy are applied to iris recognition problem. The obtained results indicate the usefulness of MPCA to select discriminative features and fuse them effectively. The experimental results reveal that the proposed tensor-based MPCA approach achieved a competitive matching performance on the SDUMLA-HMT Iris database with an adequate acceptable rate.</p>

Ranking Tensor Subspaces in Weighted Multilinear Principal Component Analysis

Multilinear principal component analysis (MPCA) has been applied for tensor decomposition and dimensionality reduction in image databases modeled through higher order tensors. Despite the well-known attractive properties of MPCA, the traditional approach does not incorporate prior information in order to steer its subspace computation. In this paper, we propose a method to explicitly incorporate such semantics in the MPCA framework to allow an automatic selective treatment of the variables that compose the patterns of interest. The method relies on spatial weights calculated, in this work, by separating hyperplanes and Fisher criterion. In this way, we can perform feature extraction and dimensionality reduction taking advantage of high level information in the form of labeled data. Besides, the corresponding tensor components are ranked in order to identify the principal weighted tensor subspaces for classification tasks. In the computational results we consider gender and facial expression experiments to illustrate the capabilities of the method for dimensionality reduction, classification and reconstruction of face images.