改进PSO优化神经网络算法的人体姿态识别

为了提高人体姿态的识别精度,提出一种基于改进的粒子群优化(PSO)神经网络的人体姿态识别算法.采用加速度传感器获取加速度信息,并在常用特征集的基础上,加入离散系数和曲线积分两种新特征作为神经网络的输入;在利用PSO神经网络参数的同时,通过控制概率,自适应地对粒子进行遗传操作,增强粒子跳出局部极小值的能力;采用训练后的神经网络对6种人体姿态进行识别.实验结果表明:该算法收敛速度和全局寻优能力得到了提高,与其他经典算法相比识别精度更高.

Human posture recognition based on neural network optimized by improved PSO algorithm

In order to improve recognition precision for human posture,an improved particle swarm optimization (PSO) algorithm neural network is proposed to recognize human posture.Acceleration sensor is used to collect acceleration information.Based on common feature sets,discrete coefficient and curve integral are added,as input of neural network,by Controlling probability,particles are operated by genetic operation when parameters of neural network are optimized by PSO algorithm and it will make the particles jump out of local the mininum value.The six postures are recognized through trained neural network.Experimental results show that the proposed algorithm can improve the convergence speed and the ability of global optimization.Compared with other classical algorithms,the proposed algorithm has higher recognition precision.