摘要
为了有效提高机器人的智能交互水平,设计一种基于深度学习的的心理情绪智能交互系统。通过分离卷积实现深度学习,对人的面部表情(喜悦、伤心、气愤、恐惧等)进行分析和特征提取,实验结果显示训练后的模型对面部表情测试集的识别准确率可达71.1%。系统分别设计了与6种不同面部表情向对应的NAO机器人肢体动作,实验结果表明,机器人可在2 s内完成识别并进行动作反馈,且连续10帧的识别结果较为准确。
In order to effectively improve the intelligent interaction of robot,an intelligent interaction and feedback system of psychology and emotion based on deep learning is proposed.Deep learning is realized by separating convolution,and human facial expressions(joy,sadness,anger,fear,etc.)are analyzed and these features are extracted.The experimental results show that the recognition accuracy of the trained model face expression test set can reach 71.1%.The limb movements of NAO robot corresponding to six different facial expressions are designed,respectively.The experimental results show that the robot can complete the recognition and action feedback within 2 s,and the recognition results of 10 consecutive frames are more accurate.
作者
张成玉
ZHANG Chengyu(Student Work Department(Student Office),Armed Forces Department,Xi’an Medical University,Xi’an 710021,China)
出处
《微型电脑应用》
2024年第2期70-73,共4页
Microcomputer Applications
基金
陕西省教育科学“十三五”规划2020年度课题(SGH20Y1458)。
关键词
深度学习
心理情绪
智能交互与反馈系统
机器人
面部表情
deep learning
psychology and emotion
intelligent interaction and feedback system
robot
facial expression
