| 基于深度学习的行为识别算法综述 |
| 投稿时间：2019-11-07 修订日期：2020-05-18 点此下载全文 |
| 引用本文： |
| 摘要点击次数: 479 |
| 全文下载次数: 0 |
|
| 基金项目:国家自然科学基金（61773219, 61701244），国家重点研发计划重点专项课题（2018YFC1405703） |
|
| 中文摘要:人体行为识别一直是计算机视觉研究中的热点，随着近几年人体行为识别在虚拟现实、短视频等方面的广泛应用，深度学习算法的快速发展，基于深度学习的行为识别算法层出不穷。相较于传统方法，基于深度学习的行为识别算法具有鲁棒性强、准确率高的优点。基于此，本文对近年来提出的基于深度学习的行为识别算法进行了梳理，并对由双流卷积网络和3D卷积网络两个网络结构发展而来的行为识别的系列算法进行了重点介绍，并总结了各个模型算法的性能和成果，最后对该领域提出了进一步的展望。 |
| 中文关键词:行为识别、深度学习、卷积网络 |
| |
| A survey of Action Recognition Algorithms based on Deep Learning |
|
|
| Abstract:Human action recognition is a hot topic in Computer Vision research.With the wide application of human action recognition in virtual reality,short video ,etc,and fast development of Deep Learning in recent years,the action recognition algorithms based on deep learning emerge in an endless stream.Compared with traditional methods,the action recognition algorithms based on deep learning have advantages of strong robustness and high accuracy.Based on this,this paper combs the action recognition algorithms based on deep learning proposed in recent years,and focuses on the series of action recognition algorithms developed from the two network structures of two-stream network and 3D convolutional network,summarizes the performance and achievements of each model algorithm,and finally puts forward further prospects in this field. |
| keywords:action recognition deep learning convolution network |
| 查看/发表评论 下载PDF阅读器 |
