Wyniki wyszukiwania - BazTech

Ograniczanie wyników

2 Optica Applicata

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: superior speed

Sortuj według:

Ogranicz wyniki do:

Fast near-infrared palmprint recognition using nonnegative matrix factorization extreme learning machine

Xu X., Zhang X., Lu L., Deng W., Zuo K

Optica Applicata

2014

Vol. 44, nr 2

285--298

Support vector machine and artificial neural network are widely used in classification applications. Extreme learning machine (ELM) is a novel and efficient learning algorithm based on the generalized single hidden layer feed forward networks, which performs well in classification applications. The research results have shown the superiority of ELM with the existing classical algorithms: support vector machine (SVM) and back propagation neural network. In this study, we firstly propose a novel nonnegative matrix factorization extreme learning machine (NMFELM) to improve the performance of standard ELM method. Then we propose a novel near-infrared palmprint recognition approach based on NMFELM classifier. As the test data, we use the near-infrared palmprint database provided by Hong Kong Polytechnic University. The experimental results demonstrate that the proposed NMFELM method outperforms the standard ELM- and SVM-based methods.

Quantum-inspired particle swarm optimization algorithm with performance evaluation of fused images

Le Z, Xinman Z., Xuebin X, Dong W., Jie L., Yang L.

Optica Applicata

2013

Vol. 43, nr 4

679--691

In order to improve and accelerate the speed of image integration, an optimal and intelligent method for multi-focus image fusion is presented in this paper. Based on particle swarm optimization and quantum theory, quantum particle swarm optimization (QPSO) intelligent search strategy is introduced in salience analysis of a contrast visual masking system, combined with the segmentation technique. The superiority of QPSO is quantum parallelism. It has stronger search ability and quicker convergence speed. When compared with other classical or novel fusion methods, several metrics for image definition are exploited to evaluate the performance of all the adopted methods objectively. Experiments are performed on both artificial multi-focus images and digital camera multi-focus images. The results show that QPSO algorithm is more efficient than non-subsampled contourlet transform, genetic algorithm, binary particle swarm optimization, etc. The simulation results demonstrate that QPSO is a satisfying image fusion method with high accuracy and high speed.