RNN循环神经网络的服务机器人交互手势辨识
Interactive Gesture Recognition of Service Robot Based on RNN Recurrent Neural Network
郑奕捷 1李翠玉 1郑祖芳1
作者信息
- 1. 湖北工业大学工业设计学院,湖北 武汉 430070
- 折叠
摘要
服务机器人交互过程中机器人重要关节点难以确定,导致交互手势辨识难以增加,因此设计一种基于RNN循环神经网络的服务机器人交互手势辨识方法.利用Kinect捕获服务机器人交互手势深度图像,确定服务机器人交互过程中的重要关节点,提取服务机器人交互手势特征.根据手势特征提取结果,定义手势模板,采用RNN循环神经网络对手势模板进行学习处理,搭建服务机器人交互手势辨识模型,得到相关的交互手势辨识结果.实验测试结果表明,采用所提方法可以快速获取高精度的服务机器人交互手势辨识结果,实际应用效果好.
Abstract
In the process of service robot interaction,it is difficult to determine the important joint points of the robot,which leads to the difficulty of increasing the interactive gesture recognition.Therefore,an interactive gesture recognition method of service ro-bot based on RNN recurrent neural network is designed.Kinectis used to capture the depth image of the interaction gestures of the service robot,determine the important joint points in the interaction process,and extract the features of the interaction gestures.Based on the results of gesture feature extraction,a gesture template is defined,and the RNN recurrent neural network is used to learn and process the gesture template.The service robot interactive gesture recognition model is built,and the relevant interactive gesture recognition results are obtained.The experimental results show that the proposed method can quickly obtain high-preci-sion interactive gesture recognition results of service robots,and the practical application effect is good.
关键词
RNN循环神经网络/服务机器人/交互手势/辨识Key words
RNN Cyclic Neural Network/Service Robot/Interactive Gestures/Identification引用本文复制引用
基金项目
教育部人文社会科学艺术青年基金(2018)(18YJC760147)
出版年
2024
NETL
NSTL
维普
万方数据