【摘 要】
:
Aiming at the problem of unsatisfactory effects of traditional micro-expression recognition algo-rithms,an efficient micro-expression recognition algorithm is proposed,which uses convolutional neu-ral networks (CNN) to extract spatial features of micro-ex
【机 构】
:
School of Electronic and Engineering,Xi’an University of Posts and Telecommunications,Xi’an 710121,P
论文部分内容阅读
Aiming at the problem of unsatisfactory effects of traditional micro-expression recognition algo-rithms,an efficient micro-expression recognition algorithm is proposed,which uses convolutional neu-ral networks (CNN) to extract spatial features of micro-expressions,and long short-term memory net-work (LSTM) to extract time domain features.CNN and LSTM are combined as the basis of micro-ex-pression recognition.In many CNN structures,the visual geometry group (VGG) using a small convo-lution kernel is finally selected as the pre-network through comparison.Due to the difficulty of deep learning training and over-fitting,the dropout method and batch normalization method are used to solve the problem in the VGG network.Two data sets CASME and CASME Ⅱ are used for test comparison,in order to solve the problem of insufficient data sets,randomly determine the starting frame,and a fixed-length frame sequence is used as the standard,and repeatedly read all sample frames of the entire data set to achieve traversal and data amplification.Finally,a high recognition rate of 67.48% is achieved.
其他文献
针对样本效果评价值为灰色异构数据的多属性群决策问题,考虑决策者主观风险的态度,提出一种基于前景理论的多属性群决策方法.首先,给出灰色异构数据的核向量和灰度向量定义,构建基于核正负理想向量与灰度正负理想向量的新的群决策参考点,综合考虑决策者意见一致性原理和综合前景值最大化准则,构建专家权重优化模型,进而确定各方案的综合前景值及排序关系;最后,通过实例分析表明该方法的有效性和实用性.
冻融作用是一种能够显著改变土的物理力学性质的强风化作用,危害寒区工程建设的安全.以兰州地区常见的黄土为样本,选取饱和渠基土样,就不同温度下试样冻胀、融沉的变化规律展开研究,观察冻融作用前后的力学性质变化.研究得到了饱和渠基土在单向冻融后,物理力学性质呈现随冷端温度变化的规律性结论.具体表现为冻胀率和融沉系数随冷端温度降低而降低;先期固结压力Pc和压缩模量Es因冻融作用减小,且影响随温度降低而降低,与干容重变化相似;粘聚力因冻融作用降低,而内摩擦角增大,该作用随冷
利用特征值法以及Hurwitz法,研究了具有年龄结构的两种群互惠模型边界平衡点的稳定性,同时给出了判断两种群互惠模型正平衡点渐近稳定的一个充分条件.
GDP时间序列数据具有线性和非线性特征.自回归模型能解释线性部分,而单纯地使用线性回归分析GDP时间序列数据不符合实际情况.将自回归模型中不具显著性的解释变量作为非参数部分引入自回归模型,充分考虑GDP时间序列的线性和非线性的混合特点,构建半参数可加自回归模型,该模型更符合现实意义.利用我国1978~2019年的GDP数据进行实证分析,结果表明该模型不仅反应了GDP时间序列的非线性特征,而且在解释能力、拟合效果和预测精准度上有一定优势.
随着对大功率、高可靠性和安全储能设备的需求不断增加,与超级电容器相关的研究数量显著增加.主要概述了超级电容器的发展史,并总结归纳了超级电容器的结构及储能机理,最后对未来储能系统的发展趋势进行了展望.
为了在三维人脸模型上实时地模拟人脸的运动,提出一种基于多媒体数据库的三维动漫人脸高真实感建模方法.将三维动漫人脸作为主要研究对象,选取特征点建立人脸特征集合.利用聚类分析和二值化处理进行三维动漫人脸区域特征点自动标定,同时采用DFFD算法对模型进行局部调整,使用立体点云数据Morphable Model表示调整后的人脸特征模型.仿真实验结果表明,所提方法能够精准标注三维动漫人脸特征点,获取理想的建
利用D-S理论对原始数据建立信度函数,划分状态区间,并对不同时期历史数据进行加权,以规范化的各阶自相关系数为权重,创建加权马尔科夫链预测模型.通过特征值法对预测对象进行定量分析,计算具体预测值.根据实例进行实证对比分析,得到了较小的预测误差.
Image super-resolution (SR) is an important technique for improving the resolution and quality of images.With the great progress of deep learning,image super-resolution achieves remarkable im-provements recently.In this work,a brief survey on recent advan
This paper investigates the networked flight control system for a laboratory 3 degrees of freedom(3-DOF) helicopter,and presents a novel networked guaranteed cost proportion-integration-differen-tiation (PID) attitude tracking control method with consider
The output current harmonic distortion of a three-level inverter is less than the traditional two-level inverter.The voltage stress of the semiconductor switch is low.A neutral point potential drift control method is proposed to solve the problem of the n