基于神经网络的不变性模式识别的人脸识别系统

Face Recognition System Based on Multilayer Neural Networks for Motion Invariant Pattern Recognition

文章提出了一种优化的多层神经网络,能完成对图像不变性模式识别。基于此多层神经网络开发了人脸识别系统。仿真实验表明该人脸识别系统,具有较高的识别速度、准确率、容错性和鲁棒性,并且基本解决了开发实用化的人脸识别系统所面临的问题,即模式识别不变性的问题。

The paper brings forward a new multilayer neural networks for motion invariant pattern recognition.A new face recognition system based on multiplayer neural network is also presented.Simulations show that it not only has upper recognition rapidity,accuracy,tolerance and rubustness,but also meet most of the demands for developing practical face recognition systems ,basically.