基于序列特征的2D CNN的动态手势识别
|
摘要
为了兼顾识别准确度和运行速度,改进了2D卷积神经网络提取多帧特征并使用长短期记忆网络进行处理特征序列,使用Softmax分类器输出分类结果;实验结果表明:基于序列特征的2D CNN网络在CHGDs数据集上的识别准确率达86. 97%,比CNN卷积神经网络提高了11. 99%,与3D CNN性能基本相当的同时,速度是3D CNN的6.98倍。
In order to balance the recognition accuracy and running speed,this paper improved 2D convolutional neural network to extract multi-frame features and used long short-term memory( LSTM) to process the feature sequences,and used the softmax classifier to output the classification results. The experimental results show that the recognition accuracy of the 2D CNN network based on sequence features on the CHGDs dataset is 86. 97%,which is 11. 99% higher than that of the CNN convolutional neural network. Compared with the performance of 3D CNN,the speed is 6. 98 of 3D CNN Times.
In order to balance the recognition accuracy and running speed,this paper improved 2D convolutional neural network to extract multi-frame features and used long short-term memory( LSTM) to process the feature sequences,and used the softmax classifier to output the classification results. The experimental results show that the recognition accuracy of the 2D CNN network based on sequence features on the CHGDs dataset is 86. 97%,which is 11. 99% higher than that of the CNN convolutional neural network. Compared with the performance of 3D CNN,the speed is 6. 98 of 3D CNN Times.
引文
[1]王海鹏,龚岩,刘武,等.一种时空多尺度适应的手势识别方法研究[J].计算机科学,2017,44(12):287-291.
[2]肖寒君.基于深度传感器的手势追踪系统设计与实现[D].哈尔滨:哈尔滨工业大学,2017.
[3]赵小川.手势识别技术研究综述与应用展望[C]//第十一届全国信号和智能信息处理与应用学术会议专刊.中国高科技产业化研究会智能信息处理产业化分会、中国高科技产业化研究会信号处理专家委员会,2017.
[4]刘蓉,刘明.基于三轴加速度传感器的手势识别[J].计算机工程,2011,37(24):141-143.
[5]BARROS P,PARISI G I,JIRAK D,et al.Real-time Gesture Recognition Using a Humanoid Robot with a Deep Neural Architecture[C]//Humanoid Robots(Humanoids),201414th IEEE-RAS International Conference on.IEEE,2014:646-651.
[6]王浩宇,漆晶,方天恩,等.基于轨迹模板匹配的动态手势识别方法[J].单片机与嵌入式系统应用,2017,17(7):39-43.
[7]郭兴伟,葛元,王林泉.基于形状特征的字母手势的分类及识别算法[J].计算机工程,2004,30(18):130-132.
[8]刘小建,张元.基于多特征提取和SVM分类的手势识别[J].计算机工程与设计,2017,38(4):953-958.
[9]LEE J S,LEE Y J,LEE E H,et al.Hand Region Extraction and Gesture Recognition from Video Stream with Complex Background through Entropy Analysis[C]//The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.IEEE,2004:1513-1516.
[10]LIU N,LOVELL B C,KOOTSOOKOS P J,et al.Model Structure Selection&Training Algorithms for an HMM Gesture Recognition System[C]//Frontiers in Handwriting Recognition,2004.IWFHR-9 2004.Ninth International Workshop on.IEEE,2004:100-105.
[11]刘江华,陈佳品,程君实.基于光流及耦合隐马尔可夫模型的动态手势识别[J].上海交通大学学报,2003,37(5):720-723.
[12]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.Imagenet Classification with Deep Convolutional Neural Networks[C]//Advances in neural information processing systems.2012:1097-1105.
[13]万鸾飞.基于多机器人离线编程系统设计[J].重庆工商大学学报(自然科学版),2016,33(6):94-97.
[14]SZEGEDY C,LIU W,JIA Y,et al.Going Deeper with Convolutions[C]//Proceedings of the IEEE conference on computer vision and pattern recognition,2015:1-9.
[15]HE K,ZHANG X,REN S,et al.Deep Residual Learning for Image Recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition,2016:770-778.
[16]JI S,XU W,YANG M,et al.3D Convolutional Neural Networks for Human Action Recognition[J].IEEE transactions on pattern analysis and machine intelligence,2013,35(1):221-231.
[17]LECUN Y,BOTTOU L,BENGIO Y,et al.Gradient-based Learning Applied to Document Recognition[J].Proceedings of the IEEE,1998,86(11):2278-2324.
[18]KIM T K,CIPOLLA R.Canonical Correlation Analysis of Video Volume Tensors for Action Categorization and Detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(8):1415-1428.
[19]易生,梁华刚,茹锋.基于多列深度3D卷积神经网络的手势识别[J].计算机工程,2017,34(8):243-248.
[2]肖寒君.基于深度传感器的手势追踪系统设计与实现[D].哈尔滨:哈尔滨工业大学,2017.
[3]赵小川.手势识别技术研究综述与应用展望[C]//第十一届全国信号和智能信息处理与应用学术会议专刊.中国高科技产业化研究会智能信息处理产业化分会、中国高科技产业化研究会信号处理专家委员会,2017.
[4]刘蓉,刘明.基于三轴加速度传感器的手势识别[J].计算机工程,2011,37(24):141-143.
[5]BARROS P,PARISI G I,JIRAK D,et al.Real-time Gesture Recognition Using a Humanoid Robot with a Deep Neural Architecture[C]//Humanoid Robots(Humanoids),201414th IEEE-RAS International Conference on.IEEE,2014:646-651.
[6]王浩宇,漆晶,方天恩,等.基于轨迹模板匹配的动态手势识别方法[J].单片机与嵌入式系统应用,2017,17(7):39-43.
[7]郭兴伟,葛元,王林泉.基于形状特征的字母手势的分类及识别算法[J].计算机工程,2004,30(18):130-132.
[8]刘小建,张元.基于多特征提取和SVM分类的手势识别[J].计算机工程与设计,2017,38(4):953-958.
[9]LEE J S,LEE Y J,LEE E H,et al.Hand Region Extraction and Gesture Recognition from Video Stream with Complex Background through Entropy Analysis[C]//The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.IEEE,2004:1513-1516.
[10]LIU N,LOVELL B C,KOOTSOOKOS P J,et al.Model Structure Selection&Training Algorithms for an HMM Gesture Recognition System[C]//Frontiers in Handwriting Recognition,2004.IWFHR-9 2004.Ninth International Workshop on.IEEE,2004:100-105.
[11]刘江华,陈佳品,程君实.基于光流及耦合隐马尔可夫模型的动态手势识别[J].上海交通大学学报,2003,37(5):720-723.
[12]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.Imagenet Classification with Deep Convolutional Neural Networks[C]//Advances in neural information processing systems.2012:1097-1105.
[13]万鸾飞.基于多机器人离线编程系统设计[J].重庆工商大学学报(自然科学版),2016,33(6):94-97.
[14]SZEGEDY C,LIU W,JIA Y,et al.Going Deeper with Convolutions[C]//Proceedings of the IEEE conference on computer vision and pattern recognition,2015:1-9.
[15]HE K,ZHANG X,REN S,et al.Deep Residual Learning for Image Recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition,2016:770-778.
[16]JI S,XU W,YANG M,et al.3D Convolutional Neural Networks for Human Action Recognition[J].IEEE transactions on pattern analysis and machine intelligence,2013,35(1):221-231.
[17]LECUN Y,BOTTOU L,BENGIO Y,et al.Gradient-based Learning Applied to Document Recognition[J].Proceedings of the IEEE,1998,86(11):2278-2324.
[18]KIM T K,CIPOLLA R.Canonical Correlation Analysis of Video Volume Tensors for Action Categorization and Detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(8):1415-1428.
[19]易生,梁华刚,茹锋.基于多列深度3D卷积神经网络的手势识别[J].计算机工程,2017,34(8):243-248.