摘要
现阶段的手势关键点特征提取在复杂背景环境下存在特征提取精度低等问题,为了解决传统方法中存在的问题,提出复杂环境下多模态手势关键点特征提取算法。首先,通过改进细菌觅食(BFO)优化算法对手势图像进行增强处理;其次,通过条件生成对抗网络对手势图像进行背景去除处理;最后,通过GIFT方法检测手势图像的关键点,并通过多尺度双树复小波变换方法和Gabor滤波方法对手势图像进行多模态手势关键点特征提取。实验结果表明:所提算法的手势关键点特征提取精度更高、效果更好。
At present,there are problems with low accuracy in feature extraction of gesture key points in complex background environments.In order to solve the problems existing in traditional methods,a multimodal gesture key point feature extraction algorithm research is proposed in complex environments.Firstly,the gesture image is enhanced by improving the bacterial foraging(BFO)optimization algorithm;Secondly,background removal is performed on gesture images through conditional generation of adversarial networks;Finally,the GIFT method is used to detect the key points of the gesture image,and the multimodal gesture key poiti-scale dual tree complex wavelet transform method and Gabor filtering method.The experimental results show that the proposed algorithm has higher accuracy and better performance in extracting gesture key point features.
作者
赖丹晖
罗伟峰
袁旭东
邱子良
LAI Dan-hui;LUO Wei-feng;YUAN Xu-dong;QIU Zi-liang(Department of Electronics and Information Engineering,The Hong Kong Polytechnic University,Hong Kong 100872,China;China Southern Power Grid Shenzhen Power Supply Bureau Co.,Ltd.,Shenzhen 518000,China)
出处
《吉林大学学报(工学版)》
EI
CAS
CSCD
北大核心
2024年第8期2288-2294,共7页
Journal of Jilin University:Engineering and Technology Edition
基金
国家自然科学基金项目(51869007)
深圳供电局有限公司南网私有云IaaS基础平台建设(先进多元智能算力平台)项目(090000HJ42210031).
关键词
改进细菌觅食优化算法
条件生成对抗网络
GABOR滤波器
双树复小波变换
关键点特征提取
improving bacterial foraging optimization algorithms
conditional generation adversarial network
gabor filter
double tree complex wavelet transform
key point feature extraction