基于堆叠去噪自编码器算法的穿墙人体检测(英文)

Through wall human detection based on stacked denoising autoencoder algorithm

超宽带雷达在穿墙人体检测中的应用已经越来越成熟,将堆叠去噪自编码器算法应用于穿墙人体状态的识别和分类中,首先使用无监督学习方法对自编码器网络进行训练,从而获得原始数据更加抽象的特征表示;然后在堆叠去噪自编码器网络的最后一层添加一个分类器.使用有监督的学习方法对网络进行微调,获得最优化的模型;最后,将测试集输入到已经训练好的网络模型上进行测试.实验结果表明,堆叠去噪自编码器深度网络可以对穿墙人体目标状态进行有效地分类识别.

The application of ultra-wideband radar in the detection of through wall human has been more mature. Thealgorithm of stacked denoising autoencoder（ SDAE）is applied to identify and classify the through wall human status. Theunsupervised learning method is used to train the autoencoder network in order to obtain more abstract feature of the originaldata, then a classifier is added at the end of the network. The supervised learning method is used to fine-tuning the network toget the optimization of the model. At last, set data is tested into the network model for testing. Experimental results showedthat the stacked denoising autoencoder deep network can effectively classify and identify the through wall human status.