循环卷积神经网络在视频动作识别中的应用

Application of Cyclic Convolution Neural Network in Video Action Recognition

使用4种类型的循环神经网络模型(RNN、GRU、LSTM、BLSTM)处理手机传感器采集的异构时间序列数据,用于人体行为识别研究。针对4种模型,分别构建自动特征提取方法,并对参数进行优化。

In this paper, four types of recurrent neural network models(RNN, GRU, LSTM, BLSTM) are used to process heterogeneous time series data collected by mobile phone sensors for human behavior recognition. For the four models, the automatic feature extraction methods are constructed and the parameters are optimized.