龚苏明
江南大学轻工过程先进控制教育部重点实验室 无锡 214122
基金项目:国家自然科学基金(61573168)
详细信息
作者简介:陈莹：女，1976年生，教授，博士，研究方向为信息融合、模式识别. Euclid
龚苏明：男，1995年生，硕士生，研究方向为计算机视觉与模式识别
通讯作者:陈莹 chenying@jiangnan.edu.cn
中图分类号:TN911.73; TP391.4计量
文章访问数:167
HTML全文浏览量:73
PDF下载量:51
被引次数:0
出版历程
收稿日期:2020-05-29
修回日期:2021-06-03
网络出版日期:2021-08-24
刊出日期:2021-12-21
Human Action Recognition Network Based on Improved Channel Attention Mechanism
Ying CHEN,,Suming GONG
Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), Jiangnan University, Wuxi 214122, China
Funds:The National Natural Science Foundation of China (61573168)
摘要
摘要:针对现有通道注意力机制对各通道信息直接全局平均池化而忽略其局部空间信息的问题，该文结合人体行为识别研究提出了两种改进通道注意力模块，即矩阵操作的时空(ST)交互模块和深度可分离卷积(DS)模块。ST模块通过卷积和维度转换操作提取各通道时空加权信息数列，经卷积得到各通道的注意权重；DS模块首先利用深度可分离卷积获取各通道局部空间信息，然后压缩通道尺寸使其具有全局的感受野，接着通过卷积操作得到各通道注意权重，进而完成通道注意力机制下的特征重标定。将改进后的注意力模块插入基础网络并在常见的人体行为识别数据集UCF101和HDBM51上进行实验分析，实现了准确率的提升。
关键词:行为识别/
通道注意力/
时空特征/
深度可分离卷积
Abstract:To tackle the problem that the existing channel attention mechanism uses global average pooling to generate channel-wise statistics while ignoring its local spatial information, two improved channel attention modules are proposed for human action recognition, namely the Spatial-Temporal (ST) interaction block of matrix operation and the Depth-wise-Separable (DS) block. The ST block extracts the spatiotemporal weighted information sequence of each channel through convolution and dimension conversion operations, and obtains the attention weight of each channel through convolution. The DS block uses firstly depth-wise separable convolution to obtain local spatial information of each channel, then compresses the channel size to make it have a global receptive field. The attention weight of each channel is obtained via convolution operation, which completes feature re-calibration with the channel attention mechanism. The proposed attention block is inserted into the basic network and experimented over the popular UCF101 and HDBM51 datasets, and the results show that the accuracy is improved.
Key words:Action recognition/
Channel attention/
Spatiotemporal feature/
Depth-wise-Separable(DS) convolution
PDF全文下载地址:
https://jeit.ac.cn/article/exportPdf?id=02daf045-008c-42d4-917c-d5869376907d
