基于粒子群算法和神经网络的人脸识别分类器研究

摘要

Because BP neural network for face recognition classifier has slow convergence and easy to fall into the local minimum, using particle swarm optimization (PSO) to improve the BP network training method, establishing an algorithm based on the improved PSO-BP neural network, which can more reasonable and effectively to determine the neural network connection weights and thresholds, applying this method to the classification of the face recognition system, and compared results with using the BP neural network classification only, experiment shows that the recognition speed is quicker and recognition result is better.%针对BP神经网络作为人脸识别分类器具有的收敛速度慢、易陷入局部极小等缺点，提出利用改进的粒子群优化算法(PSO)改善BP网络训练的方法，建立种基于改进的PSO-BP神经网络，更合理有效地确定神经网络的连接权值和阈值，将其应用到人脸识别系统中的分类环节中，并与单独使用BP神经网络分类的结果相比较，实验表明，该方法识别速度快,识别效果更好。