A VQ based efficient speech recognition method is introduced, and the key parameters of this method are comparatively studied. This method is especially designed for mandarin speaker dependent small size word set r...A VQ based efficient speech recognition method is introduced, and the key parameters of this method are comparatively studied. This method is especially designed for mandarin speaker dependent small size word set recognition. It has less complexity, less resource consumption but higher ARR (accurate recognition rate) compared with traditional HMM or NN approach. A large scale test on the task of 11 mandarin digits recognition shows that the WER(word error rate) can reach 3 86%. This method is suitable for being embedded in PDA (personal digital assistant), mobile phone and so on to perform voice controlling like digits dialing, name dialing, calculating, voice commanding, etc.展开更多
设计了一种文本相关语音锁。提取语音的梅尔频谱倒谱系数(MFCC)及其一阶差分特征,利用矢量量化算法,训练说话者的语音特征模型。通过设计实验,分析了在不同的码本数量、语音样本信噪比水平、语音特征类别组合的情况下,系统的识别效果。...设计了一种文本相关语音锁。提取语音的梅尔频谱倒谱系数(MFCC)及其一阶差分特征,利用矢量量化算法,训练说话者的语音特征模型。通过设计实验,分析了在不同的码本数量、语音样本信噪比水平、语音特征类别组合的情况下,系统的识别效果。实验结果表明:在一定的信噪比下,码本数量 N =48时,识别准确率达到95 %以上。移植到嵌入式平台后可正常运行,说明算法能够满足工程实用性要求。展开更多
针对新一代多用途视频编码(versatile video coding,VVC)标准相比上一代高效视频编码(high efficiency video coding,HEVC)采用了更多数目的时空预测模式,为相邻编码帧带来了更强的帧间相关性的问题,基于深度增强学习方法提出了一种适用...针对新一代多用途视频编码(versatile video coding,VVC)标准相比上一代高效视频编码(high efficiency video coding,HEVC)采用了更多数目的时空预测模式,为相邻编码帧带来了更强的帧间相关性的问题,基于深度增强学习方法提出了一种适用于VVC编码器的码率控制算法。首先选择合适的模型输入信息,包括帧间相关信息、分层编码结构信息和视频内容信息等;其次利用上述信息,结合长短期记忆(long short-term memory,LSTM)神经网络和增强学习方法,构建基于深度增强学习的帧间量化参数预测模型,以优化VVC编码器的码率控制过程;最后验证所提出算法的性能,将所提出算法在VTM 5.1平台实现,并与VVC源编码器进行性能对比。测试结果表明,在相同码率条件下,所提出算法相比于VVC源编码器,实现了BDBR平均节省1.81%和BDPSNR提升0.14 dB。展开更多
We obtain the quantized momentum eigenvalues, <em>P<sub>n</sub></em>, and the momentum eigenstates for the space-like Schrodinger equation, the Feinberg-Horodecki equation, with the improved de...We obtain the quantized momentum eigenvalues, <em>P<sub>n</sub></em>, and the momentum eigenstates for the space-like Schrodinger equation, the Feinberg-Horodecki equation, with the improved deformed exponential-type potential which is constructed by temporal counterpart of the spatial form of these potentials. We also plot the variations of the improved deformed exponential-type potential with its momentum eigenvalues for few quantized states against the screening parameter.展开更多
A thorough analysis of composite inertial motion (relativistic sum) within the framework of special relativity leads to the conclusion that every translational motion must be the symmetrically composite relativistic s...A thorough analysis of composite inertial motion (relativistic sum) within the framework of special relativity leads to the conclusion that every translational motion must be the symmetrically composite relativistic sum of a finite number of quanta of velocity. It is shown that the resulting spacetime geometry is Gaussian and the four-vector calculus to have its roots in the complex-number algebra. Furthermore, this results in superluminality of signals travelling at or nearly at the canonical velocity of light between rest frames even if resting to each other.展开更多
The empirically reported values of the critical current density (<i>j<sub>c</sub></i>) of Bi-2212 as 2.4 × 10<sup>5</sup> (<i>j<sub>c</sub></i><sub&g...The empirically reported values of the critical current density (<i>j<sub>c</sub></i>) of Bi-2212 as 2.4 × 10<sup>5</sup> (<i>j<sub>c</sub></i><sub>1</sub>;Sample 1) and 1.0 × 10<sup>6</sup> A/cm<sup>2</sup> (<i>j<sub>c</sub></i><sub>2</sub>;Sample 2) are intriguing because both of them correspond to the <i>same</i> values of the temperature <i>T</i> = 4.2 K and the applied magnetic field <i>H</i> = 12 × 10<sup>4</sup> G. This difference is conventionally attributed to such factors—not all of which are quantifiable—as the geometry, dimensions and the nature of dopants and the manners of preparation of the samples which cause their granular structures, grain boundaries, alignment of the grains and so on to differ. Based on the premise that the chemical potential <i>μ</i> subsumes most of these features, given herein is a novel explanation of the said results in terms of the values of <i>μ</i> of the two samples. This paper revisits the problem that was originally addressed in [Malik G.P., Varma V.S. (2020) WJCMP, 10, 53-70] in the more accurate framework of a subsequent paper [Malik G.P., Varma V.S. (2021) JSNM, 34, 1551-1561]. Besides, it distinguishes between the contributions of the electro-electron (<i>e-e</i>) and the hole-hole (<i>h-h</i>) pairs to <i>j<sub>c</sub></i>—a feature to which no heed was paid earlier. The essence of our findings is that the <i>j<sub>c</sub></i>s of the two samples differ because they are characterized by different values of the <i>primary</i> variables <i>μ<sub>i</sub></i><sub> </sub>and <img src="Edit_e1b831e9-dc51-4c3b-bd84-fa905e3e62b5.png" alt="" />, where <img src="Edit_1f775a80-30ab-447d-861f-afb4ba8fba6a.png" alt="" /> is the effective mass of a charge-carrier and <i>m<sub>e</sub></i><sub> </sub>is the free-electron mass and <i>i</i> = 1 and 2 denote Sample 1 and Sample 2, respectively. In the scenario of the charge-carriers being <i>predominantly h-h</i> pairs, the values of these parameters are estimated to be: <i>μ</i><sub>1</sub> ≈ 12.3 meV, <i>η</i><sub>1</sub> ≈ 0.58;<i>μ</i><sub>2</sub> ≈ 22.7 meV, <i>η</i><sub>2</sub> ≈ 0.94. Following from these and similar estimates when the charge-carriers are <i>e-e</i> pairs, given below for each sample are the detailed results for the values of the <i>secondary</i> variables viz. the number density of the charge-carriers and their critical velocity, the number of occupied Landau levels and the magnetic interaction parameter.展开更多
文摘A VQ based efficient speech recognition method is introduced, and the key parameters of this method are comparatively studied. This method is especially designed for mandarin speaker dependent small size word set recognition. It has less complexity, less resource consumption but higher ARR (accurate recognition rate) compared with traditional HMM or NN approach. A large scale test on the task of 11 mandarin digits recognition shows that the WER(word error rate) can reach 3 86%. This method is suitable for being embedded in PDA (personal digital assistant), mobile phone and so on to perform voice controlling like digits dialing, name dialing, calculating, voice commanding, etc.
文摘设计了一种文本相关语音锁。提取语音的梅尔频谱倒谱系数(MFCC)及其一阶差分特征,利用矢量量化算法,训练说话者的语音特征模型。通过设计实验,分析了在不同的码本数量、语音样本信噪比水平、语音特征类别组合的情况下,系统的识别效果。实验结果表明:在一定的信噪比下,码本数量 N =48时,识别准确率达到95 %以上。移植到嵌入式平台后可正常运行,说明算法能够满足工程实用性要求。
文摘针对新一代多用途视频编码(versatile video coding,VVC)标准相比上一代高效视频编码(high efficiency video coding,HEVC)采用了更多数目的时空预测模式,为相邻编码帧带来了更强的帧间相关性的问题,基于深度增强学习方法提出了一种适用于VVC编码器的码率控制算法。首先选择合适的模型输入信息,包括帧间相关信息、分层编码结构信息和视频内容信息等;其次利用上述信息,结合长短期记忆(long short-term memory,LSTM)神经网络和增强学习方法,构建基于深度增强学习的帧间量化参数预测模型,以优化VVC编码器的码率控制过程;最后验证所提出算法的性能,将所提出算法在VTM 5.1平台实现,并与VVC源编码器进行性能对比。测试结果表明,在相同码率条件下,所提出算法相比于VVC源编码器,实现了BDBR平均节省1.81%和BDPSNR提升0.14 dB。
文摘We obtain the quantized momentum eigenvalues, <em>P<sub>n</sub></em>, and the momentum eigenstates for the space-like Schrodinger equation, the Feinberg-Horodecki equation, with the improved deformed exponential-type potential which is constructed by temporal counterpart of the spatial form of these potentials. We also plot the variations of the improved deformed exponential-type potential with its momentum eigenvalues for few quantized states against the screening parameter.
文摘A thorough analysis of composite inertial motion (relativistic sum) within the framework of special relativity leads to the conclusion that every translational motion must be the symmetrically composite relativistic sum of a finite number of quanta of velocity. It is shown that the resulting spacetime geometry is Gaussian and the four-vector calculus to have its roots in the complex-number algebra. Furthermore, this results in superluminality of signals travelling at or nearly at the canonical velocity of light between rest frames even if resting to each other.
文摘The empirically reported values of the critical current density (<i>j<sub>c</sub></i>) of Bi-2212 as 2.4 × 10<sup>5</sup> (<i>j<sub>c</sub></i><sub>1</sub>;Sample 1) and 1.0 × 10<sup>6</sup> A/cm<sup>2</sup> (<i>j<sub>c</sub></i><sub>2</sub>;Sample 2) are intriguing because both of them correspond to the <i>same</i> values of the temperature <i>T</i> = 4.2 K and the applied magnetic field <i>H</i> = 12 × 10<sup>4</sup> G. This difference is conventionally attributed to such factors—not all of which are quantifiable—as the geometry, dimensions and the nature of dopants and the manners of preparation of the samples which cause their granular structures, grain boundaries, alignment of the grains and so on to differ. Based on the premise that the chemical potential <i>μ</i> subsumes most of these features, given herein is a novel explanation of the said results in terms of the values of <i>μ</i> of the two samples. This paper revisits the problem that was originally addressed in [Malik G.P., Varma V.S. (2020) WJCMP, 10, 53-70] in the more accurate framework of a subsequent paper [Malik G.P., Varma V.S. (2021) JSNM, 34, 1551-1561]. Besides, it distinguishes between the contributions of the electro-electron (<i>e-e</i>) and the hole-hole (<i>h-h</i>) pairs to <i>j<sub>c</sub></i>—a feature to which no heed was paid earlier. The essence of our findings is that the <i>j<sub>c</sub></i>s of the two samples differ because they are characterized by different values of the <i>primary</i> variables <i>μ<sub>i</sub></i><sub> </sub>and <img src="Edit_e1b831e9-dc51-4c3b-bd84-fa905e3e62b5.png" alt="" />, where <img src="Edit_1f775a80-30ab-447d-861f-afb4ba8fba6a.png" alt="" /> is the effective mass of a charge-carrier and <i>m<sub>e</sub></i><sub> </sub>is the free-electron mass and <i>i</i> = 1 and 2 denote Sample 1 and Sample 2, respectively. In the scenario of the charge-carriers being <i>predominantly h-h</i> pairs, the values of these parameters are estimated to be: <i>μ</i><sub>1</sub> ≈ 12.3 meV, <i>η</i><sub>1</sub> ≈ 0.58;<i>μ</i><sub>2</sub> ≈ 22.7 meV, <i>η</i><sub>2</sub> ≈ 0.94. Following from these and similar estimates when the charge-carriers are <i>e-e</i> pairs, given below for each sample are the detailed results for the values of the <i>secondary</i> variables viz. the number density of the charge-carriers and their critical velocity, the number of occupied Landau levels and the magnetic interaction parameter.
