基于脉冲神经网络人脸特征提取

Face Features Extraction Base on Spiking Neural Networks

运用基于复杂度和最佳阈值算法对人脸图像进行人眼特征定位并标准化图像,引入生物并行机制的脉冲神经网络训练输入图像,获得稳定的神经元突触强度矩阵,选取此矩阵系数作为人脸特征向量,用最近邻法则分类识别.利用该突触强度分布矩阵,注入刺激电流,神经网络中原始图像得以重建.实验证明,该方法在表情、姿态变化以及深度旋转的图像中特征定位准确,识别率较高.

A algorithm based on complexity and best threshold is used to locate human eyes and normalize images. Spiking neural networks, which inherit the parallel mechanism from biological system, are used to obtain the intensity of synapses matrix, the matrix coefficients are face features, The nearest neighbor classifier is used for matching faces. Spiking neural networks can remember key features of a visual image through synapse strength distribution and recall the visual im- age by triggering a specific neuron. Experimental results show that the proposed algorithm of eye location works well in multi-position and complex background, and the face features extraction based on spiking neural networks can achieve high recognition rate.