首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献,搜索用时 109 毫秒
1.
This work demonstrates the use of the nonlinear time-frequency distribution (NL TFD) of a discrete time energy operator (DTEO) based on amplitude modulation-frequency modulation demodulation techniques as a feature in speech recognition. The duration distribution based hidden Markov module in a speaker independent large vocabulary mandarin speech recognition system was reconstructed from the feature vectors in the front-end detection stage. The goal was to improve the performance of the existing system by combining new features to the baseline feature vector. This paper also deals with errors associated with using a pre-emphasis filter in the front end processing of the present scheme, which causes an increase in the noise energy at high frequencies above 4 kHz and in some cases degrades the recognition accuracy. The experimental results show that eliminating the pre-emphasis filters from the pre-processlng stage and using NL TFD with compensated DTEO combined with Mel frequency cepstrum components give a 21.95% reduction in the relative error rate compared to the conventional technique with 25 candidates used in the test.  相似文献   

2.
Multidimensional grey relation projection value can be synthesized as one-dimensional projection value by using projection pursuit model. The larger the projection value is, the better the model. Thus, according to the projection value, the best one can be chosen from the model aggregation. Because projection pursuit modeling based on accelerating genetic algorithm can simplify the implementation procedure of the projection pursuit technique and overcome its complex calculation as well as the difficulty in implementing its program, a new method can be obtained for choosing the best grey relation projection model based on the projection pursuit technique.  相似文献   

3.
Multi-source information can be utilized collaboratively to improve the performance of information retrieval. To make full use of the document and collection information, this paper introduces a new information retrieval model that relies on the Dempster-Shafer theory of evidence. Each query-document pair is taken as a piece of evidence for the relevance between a document and a query. The evidence is combined using Dempster’s rule of combination, and the belief committed to the relevance is obtained. Retrieved documents are then ranked according to the belief committed to the relevance. Several basic probability assignments are also proposed. Extensive experiments over the Text REtrieval Conference (TREC) test collection ClueWeb09 show that the proposed model provides performance similar to that of the Vector Space Model (VSM). Under certain probability assignments, the proposed model outperforms the VSM by 63% in terms of mean average precision.  相似文献   

4.
This paper reviews the development course of the front crossbeam assembly for a self-owned brand vehicle model based on lightweight and passive safety performance.Combining with an A00 model variant,the paper details the design of extruded aluminum-alloy front crossbeam assembly from the perspectives of optimal design,performance verification, lightweight effect and cost control.The following results in the technical and engineering applications have been achieved.The weight of the developed aluminum-alloy crossbeam can be reduced by 51%.The simulated analysis of the collision rigid wall,the 40%offset hammering as well as the static crush test of energy-absorbing box show that alter reasonable materials matching and size optimization of the crossbeam and the energy-absorbing boxes,the level of crash safety can be improved.The price of aluminum-alloy front crossbeam can be lowered by using the extruding die instead of the stamping die to reduce the die cost-sharing.  相似文献   

5.
This paper describes a parallel computing platform using the existing facilities for the digital watershed model. In this paper, distributed multi-layered structure is applied to the computer cluster system, and the MPI-2 is adopted as a mature parallel programming standard. An agent is introduced which makes it possible to be multi-level fault-tolerant in software development. The communication protocol based on checkpointing and rollback recovery mechanism can realize the transaction reprocessing. Compared with conventional platform, the new system is able to make better use of the computing resource. Experimental results show the speedup ratio of the platform is almost 4 times as that of the conventional one, which demonstrates the high efficiency and good performance of the new approach.  相似文献   

6.
A new instrument based on light projection and image analysis techniques for assessing fabric pilling is presented. This system can automatically detect the pills from the successive sections of the fabric surface, which can make up for the limitation of both the gray level image -analysis techniques and laser triangulab'on techniques. The test of a large number of worsted wool fabric samples shows that it can objectively characterize the degree of pilling by using fuzzy sets and has good accordance with human visual inspection.  相似文献   

7.
This paper presents a novel hierarchy cache architecture for the purpose of optimizing IO performance. The main idea of the hierarchy cache is to use a few megabytes of RAM and a pagefile to form a two-level cache architecture. The pagefile is equivalent to the cache disk in DCD(Disk Caching Disk). The pagefile outperforms data disks, because data are accessed in different units and different ways. Small writes are collected in the RAM cache first, and data will be transferred to the pagefile in large writes later. When the system is idle, it will destage data from the pagefile to data disks. The performance test results show that the hierarchy cache can improve IO performance dramatically for small writes, and the mail server using the hierarchy cache driver can handle transactions about 2.2 times faster than the normal mail server. The hierarchy cache is implemented as a filter driver, so it‘s transparent to the current Windows 2000/Windows XP operating system.  相似文献   

8.
Weighted geometric evaluation approach based on Projection pursuit (PP) model is presented in this paper to optimize the choice of schemes. By using PP model, the multi-dimension evaluation index values of schemes can be synthesized into projection value with one dimension. The scheme with a bigger projection value is much better, so the schemes sample can be an optimized choice according to the projection value of each scheme. The modeling of PP based on accelerating genetic algorithm can predigest the realized process of projection pursuit technique, can overcome the shortcomings of large computation amount and the difficulty of computer programming in traditional projection pursuit methods, and can give a new method for application of projection pursuit technique to optimize choice of schemes by using weighted geometric evaluation. The analysis of an applied sample shows that applying PP model driven directly by samples data to optimize choice of schemes is both simple and feasible, that its projection values are relatively decentralized and profit decision-making, that its applicability and maneuverability are high. It can avoid the shortcoming of subjective weighing method, and its results are scientific and objective.  相似文献   

9.
In order to improve the concurrent access performance of the web-based spatial computing system in cluster,a parallel scheduling strategy based on the multi-core environment is proposed,which includes two levels of parallel processing mechanisms.One is that it can evenly allocate tasks to each server node in the cluster and the other is that it can implement the load balancing inside a server node.Based on the strategy,a new web-based spatial computing model is designed in this paper,in which,a task response ratio calculation method,a request queue buffer mechanism and a thread scheduling strategy are focused on.Experimental results show that the new model can fully use the multi-core computing advantage of each server node in the concurrent access environment and improve the average hits per second,average I/O Hits,CPU utilization and throughput.Using speed-up ratio to analyze the traditional model and the new one,the result shows that the new model has the best performance.The performance of the multi-core server nodes in the cluster is optimized; the resource utilization and the parallel processing capabilities are enhanced.The more CPU cores you have,the higher parallel processing capabilities will be obtained.  相似文献   

10.
11.
基于GMM全统计参数和SVM的文本无关话者确认   总被引:1,自引:0,他引:1  
针对电话语音条件下的文本无关话者确认,提出了一种基于GMM(Gaussian mixturemodel)全统计参数和SVM(support vector machine)的话者确认方法,以克服语音特征参数直接建立SVM话者模型面临的困难.该方法使用由GM(general model)自适应均值得到的GMM提取统计参数,定义了一种合理利用全部统计参数的特征参数,并以此参数及线性核函数建立了具有良好性能的SVM话者模型.与GMM-UBM方法及另一种基于GMM统计参数和SVM的方法在NIST05SRE数据库中的实验比较,结果表明基于GMM全统计参数和SVM的话者确认方法拥有优异的性能.  相似文献   

12.
在说话人识别系统中,训练语音与测试语音的话机类型失配会使说话人识别系统识别性能显著下降。为了提高说话人识别系统的稳健性,在说话人模型合成和话机归一化的基础上提出一种新的信道补偿方法HNSSM(handsetnormalizationinsynthesizedspeakmodel),综合模型和分数两个方面对系统进行信道补偿。1999年美国国家标准技术局说话人识别评测语音库上的实验表明,采用新的信道补偿方法使系统在等错误率和最小检测代价上比仅采用倒谱均值减的基线系统分别降低了39.4%和20.9%,而且优于只采用说话人模型合成或话机归一化补偿的系统。  相似文献   

13.
14.
短时声纹识别中语言属性不匹配问题降低了话者识别系统性能,语言属性失配在短时话者识别中的影响成为当前研究热点。从消除特征参数中的语言属性信息出发,提出了一种语言属性映射方法。采用均值超向量来构建训练样本,应用高斯混合模型对每一句话不等长的特征进行建模,引入加权矩阵得到语言属性空间,通过映射方法在统计参数超矢量空间中消除语言属性的影响。使用当前两套经典声纹识别系统作为基线系统做对比试验,结果表明:相对基线系统在等错误率上,男声测试部分降低了23.07%,女声测试部分降低了20.39%,从而验证了正确性和有效性。  相似文献   

15.
针对电话手机语音的文本无关说话人确认中,训练集语音和测试集语音来自不同信道所产生失配而导致系统性能下降的问题,采用一种基于高维空间映射的方法对系统进行补偿.在分析了已有的说话人确认系统的基础上,提出了一种基于特征参数映射支持向量机模型(PSVM)的说话人确认系统.首先用大量已知信道类型的语音训练出信道空间及映射矩阵,然后训练语音和测试语音都通过映射,消除因信道不同而导致的失配影响.在NIST数据库上的实验结果表明,这种方法弥补了训练语音和测试语音的失配,说话人确认系统的性能有了明显的提高.  相似文献   

16.
针对说话人确认系统中GMM超向量建模计算复杂度高以及易受信道干扰的问题,提出一种新型的基于Bhattacharyya距离聚类的WCCN序列核函数算法.首先计算话者GMM模型之间的Bhattacharyya距离,根据该Bhattacharyya距离对话者模型进行聚类,得到聚类中心模型;紧接着对聚类中心模型的均值向量进行MAP自适应,进而生成超向量序列核函数;最后采用WCCN平滑归一化技术对序列核函数进行信道补偿,抑制噪音和信道畸变对核函数的影响.将该Bhattacharyya聚类WCCN核函数应用到SVM说话人确认系统,仿真实验结果表明该核函数可以有效地提高系统的识别准确率和识别速度.  相似文献   

17.
提出了一种新的注册者模型——“时间一空间分布模型(TSDM)”,传统GMM用参数表征训练矢量的空间分布,但舍弃了训练矢量间的时间联系信息,而TSDM利用基于GMM中均值矢量的高阶协方差矩阵,可向训练矢量的空间分布模型中引入一定程度的训练矢量间时间联系.还给出了TSDM的判据生成方法.实验表明,TSDM能在长训练语句时获得与传统GMM相当的识别性能,在短训练语句时表现得更优秀.  相似文献   

18.
提出了一种用于文本相关说说话人确认技术的i-向量提取方法和L-向量表示.一段用于注册或识别的语音可以用i-向量和L-向量联合表示.同时提出了一种改进的用于支持向量机(SVM)后端分类的核函数,改进的核函数可以同时区分说话人身份的差异和文本内容的差异.在RSR 2015语料集合1和集合2上验证系统的性能,实验结果显示改进的算法相对于传统的i-向量系统的基线能提高至多30%的识别率.  相似文献   

19.
与文本无关的复合策略说话人辨识系统   总被引:4,自引:1,他引:4  
为获得较高的说话人辨识正确率,同时减小辨识系统的时空开销,提出了一种复合策略的辨识系统。采用长时平均频谱作为粗识的特征,定义了相应的辨识判别准则。建立mel-倒谱特征的高斯混合模型(GMM)进行第二步辨识。给出了GMM求解算法的一种简便推导,着重研究了判别阈值,预加重系数,GMM阶次,训练语音长度及辨识语音长度对系统辨识性能的影响。  相似文献   

20.
支持向量机(SVM)是以统计学习理论为基础,解决模式识别问题的有力工具,但是它训练算法复杂,难以处理大量样本,限制了其在说话人识别方面的使用。针对这个问题,提出了一种基于GMM(高斯混合模型)统计参数和SVM的说话人辨认系统,以GMM模型的统计参数来训练SVM说话人辨认模型,有效解决了大样本数据下SVM模型的训练问题。实验表明,该方法有良好的效果,并且与倒谱加权方法结合后,可以增强系统的健壮性,进一步提高系统的识别率。  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号