摘要
为实现非接触式人体尺寸自动测量,通过二维人体照片中提取的二维尺寸信息来预测人体的围度信息。提出一种基于改进WOA-ENN-MC的人体围度尺寸预测模型。通过优化收敛因子、联合最优惯性权值和阿基米德螺旋更新策略优化鲸鱼算法以获得更优的寻优广度和精度,将IWOA输出的最优值作为Elman神经网络(ENN)的初始权值阈值训练,避免了传统模型在迭代后期陷入局部最优的同时缩减了网络训练的收敛时间。采用马尔可夫残差模型(MC)对IWOA-ENN模型的预测值进行优化,进一步提高模型的预测精度。对CAESAR数据集中270组样本的实测值与预测数据进行对比分析,结果表明:IWOA-ENN-MC相对于GA-ENN、PSO-ENN和WOA-ENN模型在寻优能力方面有着更明显的优势,并且模型的拟合和预测精度效果更好。
In order to realize non-contact automatic measurement of human body size and predict the human body girths through 2D size information extracted form two-dimensional body images,a human body girths predicting model based on improved WOA-ENN-MC is proposed.The better optimization breadth and accuracy were obtained by optimizing the WOA with the optimized convergence factor,the joint optimal inertia weight and the Archimedes spiral updating strategy.The optimal output of improved WOA(IWOA)was used as the initial weights and thresholds of Elman neural network(ENN).It avoided the classical model falling into local optimum in the late iteration and reduced the convergence time of network training.The Markov residual model was used to optimize the predicted values of IWOA-ENN model,which further improved the prediction accuracy of the model.The measured values of 270 samples from the CAESAR database were compared with the predicted values.The results prove that IWOA-ENN-MC model has more obvious advantages in optimization capability and accuracy of fitting and prediction than GA-ENN,PSO-ENN and WOA-ENN.
作者
胡新荣
刘嘉文
Hu Xinrong;Liu Jiawen(Engineering Research Center of Hubei Province for Clothing Information,Wuhan 430200,Hubei,China;School of Mathematics&Computer Science,Wuhan Textile University,Wuhan 430200,Hubei,China)
出处
《计算机应用与软件》
北大核心
2023年第5期190-199,共10页
Computer Applications and Software
基金
国家自然科学基金项目(61103085)
湖北省高等学校优秀中青年科技创新团队计划项目(T201807)。
