摘要
为降低步态识别特征噪声、提高BP神经网络在步态识别中的准确性和高效性,提出一种基于粒子群优化的BP神经网络识别算法。该算法将形态学细化思想融入人体骨架图特征值提取中,在二维平面上抽取多种特征值,然后建立粒子群优化神经网络模型,将特征值矩阵代入模型中,在反复迭代后产生最优迭代函数作为神经网络优化函数,不断优化网络层之间的权值和阈值。实验结果表明,优化后的模型识别率高达97.125%。
In order to reduce the characteristic noise of gait recognition and improve the accuracy and efficiency of BP neural network in gait recognition, a BP neural network recognition algorithm based on particle swarm optimization is proposed. This algorithm integrates morphological thinning into the feature extraction of human skeleton map, extracts many kinds of eigenvalues on two-dimensional plane, establishes a particle swarm optimization neural network model, substitutes eigenvalue matrix into the model, generates the optimal iteration function as the optimization function of neural network after repeated iterations, and continuously optimizes the weights and thresholds between network layers. The experimental results show that the recognition rate of the optimized model is as high as 97.125%.
作者
邹倩颖
王小芳
ZOU Qianying;WANG Xiaofang(Chengdu College of University of Electronic Science and Technology of China, Chengdu 611731, China;School of Computer Science, China West Normal University, Nanchong 637002, China)
出处
《实验技术与管理》
CAS
北大核心
2019年第8期130-133,138,共5页
Experimental Technology and Management
基金
成都市科技局重点研发支撑计划技术创新研发项目(2018-YFYF-00191-SN)
关键词
步态识别
神经网络
粒子群优化
仿真
gait recognition
neural network
particle swarm optimization
simulation
