点击次数：

论文类型：期刊论文

发表时间：2011-09-01

发表刊物：International Journal of Digital Content Technology and its Applications

收录刊物：Scopus、EI

卷号：5

期号：9

页面范围：314-322

ISSN号：19759339

摘要：This paper proposed a novel approach for face recognition based on sparse representation and recurrent neural network (RNN) optimization. In the work, the dictionary for sparse representation was constructed by using the Gabor local features of the training images to enhance the sparseness of representation coefficient vector, and a one-layer RNN model was introduced to solve the sparse representation coefficients for the purpose of real-time processing. Since the RNN model can be implemented by very large scale integrated (VLSI) circuits with parallel and distributed processing capability, the real-time processing for the sparse coding can be achieved. Simulation results are given to demonstrate the effectiveness and efficiency of the proposed method.