Narrow Results

An illumination and expression invariant face recognition method based on uniform local binary patterns (uLBP) and Legendre moments is proposed in this work. The proposed method exploits uLBP texture features and Legendre moments to make a feature representation with enhanced discriminating power. The input images are preprocessed to extract the face region and normalized. From normalized image, uLBP codes are extracted to obtain texture image which overcomes the effect of monotonic temperature changes. Legendre moments are computed from this texture image to get the required feature vector. Legendre moments conserve the spatial structure information of the texture image. The resultant feature vector is classified using k-nearest neighbor classifier with $L_1$ norm. To evaluate the proposed method, experiments are performed on IRIS and NVIE databases. The proposed method is tested on both visible and infrared images under different illumination and expression variations and performance is compared with recently published methods in terms of recognition rate, recall, length of feature vector, and computational time. The proposed method gives better recognition rates and outperforms other recent face recognition methods.

In recent years, the singular spectrum analysis (SSA) technique has been further developed and increasingly applied to solve many practical problems. The aim of this research is to introduce a new version of SSA based on $L_1$-norm. The performance of the proposed approach is assessed by applying it to various real and simulated time series, especially with outliers. The results are compared with those obtained using the basic version of SSA which is based on the Frobenius norm or $L_2$-norm. Different criteria are also examined including reconstruction errors and forecasting performances. The theoretical and empirical results confirm that SSA based on $L_1$-norm can provide better reconstruction and forecasts in comparison to basic SSA when faced with time series which are polluted by outliers.

The graph regularized nonnegative matrix factorization (GNMF) algorithm has received extensive attention in the field of machine learning. GNMF generally uses the square loss method to measure the quality of reconstructed data. However, noise is introduced when high-dimensional data is mapped to low-dimensional space, which leads to a decrease in model clustering accuracy since the square loss method is sensitive to noise. To solve this issue, this paper proposes a novel graph regularized sparse NMF (GSNMF) algorithm. For obtaining cleaner data matrices to approximate the high-dimensional matrix, the l₁-norm on the reconstructed low-dimensional matrix is added to achieve the adjustment of the data eigenvalues in the matrices and the sparse constraints of the objective function. To address the optimization process of our algorithm, the corresponding reasoning is given with an iterative updating algorithm. Experimental results on 8 datasets have shown that the proposed algorithm has a superior performance.