针对中文儿童语音情感识别的准确性问题,提出了一种结合深度卷积神经网络(Deep Convolutional Neural Network,DPCNN)与堆叠长短时记忆(Stacked Long Short Term Memory,SLSTM)网络的融合模型,旨在提高中文儿童语音情感识别的准确性。通...针对中文儿童语音情感识别的准确性问题,提出了一种结合深度卷积神经网络(Deep Convolutional Neural Network,DPCNN)与堆叠长短时记忆(Stacked Long Short Term Memory,SLSTM)网络的融合模型,旨在提高中文儿童语音情感识别的准确性。通过DPCNN对语音信号中的长距离依赖关系进行提取,再利用SLSTM捕捉情感相关的序列依赖信息,最终通过softmax分类器实现情感状态的判别。实验结果显示,基于DPCNN-SLSTM的模型在中文儿童语音数据集上的情感识别准确率达到了92%,显著优于CNN、LSTM和CNN-LSTM模型。研究结果对于推动儿童语音情感识别技术的发展具有重要意义。展开更多
老年人因年龄增长、身体机能衰退和认知功能减弱而面临不同程度的生活危险。因此,为了及时发现、监测和处理老年人的危险姿势,从而保护老年人的安全和健康。研究提出一种融合端对端思想和卷积神经网络(Port to port convo-lutional neur...老年人因年龄增长、身体机能衰退和认知功能减弱而面临不同程度的生活危险。因此,为了及时发现、监测和处理老年人的危险姿势,从而保护老年人的安全和健康。研究提出一种融合端对端思想和卷积神经网络(Port to port convo-lutional neural network,PTP-CNN)的老年人危险位姿虚拟模型识别算法,从而做出预防性措施或及时的护理。研究结果表明,该系统在运用PTP-CNN算法时,Epochs的训练次数为15~30之间,MSE评价指标上PTP-CNN模型分别比SW-CNN、AlexNet降低25.33%、5.17%,说明PTP-CNN模型拥有更高的准确性和精确性,可以更好地进行图像识别任务,从而及时发现老年人的危险姿势。展开更多
This paper focuses on the stability analysis of nonlinear networked control system with integral quadratic constraints(IQC) performance, dynamic quantization, variable sampling intervals, and communication delays. By ...This paper focuses on the stability analysis of nonlinear networked control system with integral quadratic constraints(IQC) performance, dynamic quantization, variable sampling intervals, and communication delays. By using input-delay and parallel distributed compensation(PDC) techniques, we establish the Takagi-Sugeno(T-S) fuzzy model for the system, in which the sampling period of the sampler and signal transmission delay are transformed to the refreshing interval of a zero-order holder(ZOH). By the appropriate Lyapunov-Krasovskii-based methods, a delay-dependent criterion is derived to ensure the asymptotic stability for the system with IQC performance via the H∞ state feedback control. The efficiency of the method is illustrated on a simulation exampler.展开更多
文摘针对中文儿童语音情感识别的准确性问题,提出了一种结合深度卷积神经网络(Deep Convolutional Neural Network,DPCNN)与堆叠长短时记忆(Stacked Long Short Term Memory,SLSTM)网络的融合模型,旨在提高中文儿童语音情感识别的准确性。通过DPCNN对语音信号中的长距离依赖关系进行提取,再利用SLSTM捕捉情感相关的序列依赖信息,最终通过softmax分类器实现情感状态的判别。实验结果显示,基于DPCNN-SLSTM的模型在中文儿童语音数据集上的情感识别准确率达到了92%,显著优于CNN、LSTM和CNN-LSTM模型。研究结果对于推动儿童语音情感识别技术的发展具有重要意义。
文摘老年人因年龄增长、身体机能衰退和认知功能减弱而面临不同程度的生活危险。因此,为了及时发现、监测和处理老年人的危险姿势,从而保护老年人的安全和健康。研究提出一种融合端对端思想和卷积神经网络(Port to port convo-lutional neural network,PTP-CNN)的老年人危险位姿虚拟模型识别算法,从而做出预防性措施或及时的护理。研究结果表明,该系统在运用PTP-CNN算法时,Epochs的训练次数为15~30之间,MSE评价指标上PTP-CNN模型分别比SW-CNN、AlexNet降低25.33%、5.17%,说明PTP-CNN模型拥有更高的准确性和精确性,可以更好地进行图像识别任务,从而及时发现老年人的危险姿势。
基金Supported by the National Natural Science Foundation of China(61472136)the Best Youth of the Education Department of Hunan Province(16B023)
文摘This paper focuses on the stability analysis of nonlinear networked control system with integral quadratic constraints(IQC) performance, dynamic quantization, variable sampling intervals, and communication delays. By using input-delay and parallel distributed compensation(PDC) techniques, we establish the Takagi-Sugeno(T-S) fuzzy model for the system, in which the sampling period of the sampler and signal transmission delay are transformed to the refreshing interval of a zero-order holder(ZOH). By the appropriate Lyapunov-Krasovskii-based methods, a delay-dependent criterion is derived to ensure the asymptotic stability for the system with IQC performance via the H∞ state feedback control. The efficiency of the method is illustrated on a simulation exampler.
