助手标题  
全文文献 工具书 数字 学术定义 翻译助手 学术趋势 更多
查询帮助
意见反馈
   识别输出 的翻译结果: 查询用时:0.369秒
图标索引 在分类学科中查询
所有学科
更多类别查询

图标索引 历史查询
 

识别输出
相关语句
  final recognition
     According to this algorithm,the output from CHMM is taken as the feature vector for recognition and the pre-recognition vector is extracted from the recognition vector space by the ANN,afterwards,the pre-recognition result is further processed to accomplish the final recognition.
     该方法以CHMM的输出作为系统的识别矢量,利用人工神经网络的模式分类和自学习功能,从识别矢量空间中提取语音预识别矢量,再由识别矢量对预识别结果进行识别输出
短句来源
  “识别输出”译为未确定词的双语例句
     Combination of Multiple Predictors for Correct-Incorrect Classification of Output Words in Real Time Continuous Speech Recognition
     多预测子融合实时连续语音识别输出词正误判别
短句来源
     First the 3-D scattering model of the targets is built based on the theory of radar target scattering center in the optics region and measure data,then by comparing the profiles of actual target with the profiles of model target,the two most resemble profiles are regarded as the target recognition result.
     该算法首先从光学区雷达目标的多散射中心理论出发,在大量实测(暗室或外场测量)数据的基础上,建立待识别目标的三维电磁散射模型。 目标识别时,用目标的实测距离像与数据库中所有目标的三维散射模型在任意姿态角上的投影形成的距离像进行匹配,取相关性最强的目标类型作为识别输出结果。
短句来源
     esolved brain wave is input into neural network for identification. The output results in the identification of epileptic wave.
     把分解后的脑电信号输入神经网络进行识别,输出异常波的识别结果。
短句来源
     The results show that the present method is feasible and it has advantages of few requirements of historical data, automatic increase of identification category, and the noiseproof ability.
     实验证明这种方法使网络抗噪声能力增强,使损伤识别的效果更好。 ART2网络具有自动从环境中学习的能力,能自动的给出新的识别输出
短句来源
     Under the decoding strategy of using stack decoding to rescore the word trellis to generate final output,this paper uses decision tree to combine multiple predictors to identify each of recognition output words as correct or incorrect.
     本文在采用堆栈译码词网重估输出作为识别最终输出的连续语音识别实时解码条件下,利用决策树方法将多个预测子融合,对识别输出词进行正确和错误的判别。
短句来源
更多       
  相似匹配句对
     DISTRIBUTION RECOGNITION OF SIMULATION OUTPUT DATA BASED ON NEURAL NETWORK
     基于神经网络的仿真输出数据的分布识别
短句来源
     recognition is finished.
     进行识别
短句来源
     Facial expression recognition.
     表情识别
短句来源
     A Method of Structural Modal Parameters Identification Under Unknown Inputs Using Neural Networks
     基于输出响应的识别结构参数的神经网络方法
短句来源
     output of the model.
     模型输出
短句来源
查询“识别输出”译词为用户自定义的双语例句

    我想查看译文中含有:的双语例句
例句
为了更好的帮助您理解掌握查询词或其译词在地道英语中的实际用法,我们为您准备了出自英文原文的大量英语例句,供您参考。
  final recognition
Considerable improvement has been observed, and the final recognition rate of this system ranges from 95.54 to 99.11% with a reliability of 99.54 to 99.11%.
      
The final recognition system consists of a cascade association of small MLPs, which allows minimization of the overall recognition time while retaining a high recognition rate.
      
The final recognition is achieved by employing a number of 2-class classifiers based on the Support Vector Machine (SVM) method.
      
Final recognition is accomplished by a secondary stage which performs local analysis on the characters in each primary category.
      
After all sets, a brief distractor task and final recognition task using both old and new images were com pleted.
      
更多          


An investigation on a special electronic circuit for measurement of mag ic fields with hall cell is presented in this paper. As the carrier wave lows automatically the slow change of modulating wave, so the range modulating wave amplitude which can be measured is expended. In addion,influence of slowly changing background is almost entirely eliminatd automatic gain contol with proper time constant. Therefore signals in slowly changing high background can be measured. Signal with amplitude om 50 μV_(pp) to 5mV_(pp)...

An investigation on a special electronic circuit for measurement of mag ic fields with hall cell is presented in this paper. As the carrier wave lows automatically the slow change of modulating wave, so the range modulating wave amplitude which can be measured is expended. In addion,influence of slowly changing background is almost entirely eliminatd automatic gain contol with proper time constant. Therefore signals in slowly changing high background can be measured. Signal with amplitude om 50 μV_(pp) to 5mV_(pp) can be amplifield and the output is big enough for A/D converter. When the signal is 50μV_(pp), the S/N ratio is better than3. In order to be capable of working in serious environments, all of the circuits are powered with two 1.5 volt cells. At maximum output, the power current is 10 mA. The circuit is shielded to avoid interference from outside nd between parts of the circut. It was found that the shielding is of great mportance for the carrier to follow the modulating signal amplitude change.

本文是为有关科研项目所作的研究成果.由于研究出一种自动跟踪信号幅度缓慢波动的调制电路,使检测电路对信号随机适应范围得到了较大扩展.利用变载波控制与自动增益控制时间常数有机配合,基本上消除慢波动基底信号的影响,能检测出附着在大本底上的非电转换来的信号波包,在50μv_(pp)到5mv_(pp)输入范围内都可给出A/D转换器可识别的输出幅值:50μv_(pp)输入时仍保持有不小于3的信噪比.

We describe a speaker-dependent topic-constrained continuous Chinese speech-understanding system which enables its user to obtain timetable and fare information through a spoken dialogue. A method is described which understands the results of speech recognition and generates the response to be used by the synthesiser using context-free grammer (CFG), context-senstive gramar (CSG) and blackboard. Stack technique is used to process insufficient information and information confirmation. Sematics processing obtains...

We describe a speaker-dependent topic-constrained continuous Chinese speech-understanding system which enables its user to obtain timetable and fare information through a spoken dialogue. A method is described which understands the results of speech recognition and generates the response to be used by the synthesiser using context-free grammer (CFG), context-senstive gramar (CSG) and blackboard. Stack technique is used to process insufficient information and information confirmation. Sematics processing obtains verb and other key information, so that this system has fault-tolerant ability, We describe the time using fuzzy mathematics method.

本文描述了一个以火车售票和信息查询事项为对话主题的专人使用的连续汉语人机对话系统原型,着重讨论语言理解和应答文生成,这部分是连接语音识别和语音合成的纽带,起着承上启下的作用。语言理解和应答文生成采用上下文无关文法(CFG),结合黑板以及局部上下文相关(CSG),对连续汉语语音识别输出的结果进行理解,生成人机对话系统中的应答文,作为语音合成的输入。利用堆栈技术处理不完全信息和信息证实,使得该系统具有自动提问获取知识的能力和对关键信息进行确认的能力,从而使人机对话能连续进行。语义处理提取动词等关键信息,使该系统具有容错性。我们对时间的描述采用模糊数学方法处理,以适应现实对话中不精确的时间表示。

In this study,effects of speech input and output technology on perfor-mance of man-machine systems were evaluated through four types of target tracking tasksdenoted as SK,CK,SS and CS where the first letter/S/or/C/represented the displaymodes in speech or character on CRT screen for informing the directions of tracking,and thesecond letter/S/or/K/referred to data input mode using speech or keyboard respectively.It was shown that the mean ORT(Operation/Response Time)for Task SK was 6.3% lessthan that for Task...

In this study,effects of speech input and output technology on perfor-mance of man-machine systems were evaluated through four types of target tracking tasksdenoted as SK,CK,SS and CS where the first letter/S/or/C/represented the displaymodes in speech or character on CRT screen for informing the directions of tracking,and thesecond letter/S/or/K/referred to data input mode using speech or keyboard respectively.It was shown that the mean ORT(Operation/Response Time)for Task SK was 6.3% lessthan that for Task CK,the mean ORT for Task SS was 5.8% less than that for Task CS.Among the mean ORTs for all the four types of tasks,ORT for task SS was the shortest,while that for task CK was the longest.The ORER(Operation/Response Error Rate)ofspeech display mode(2.75%)was much lower than that of CRT display mode(5.38%),however,the ORER of speech input mode(2.98%)was higher than that of keyboard in-put mode(1.13%).This might be resulted from limited performance of the speech recogni-tion system and the speech variations of individual subjects.These results imply that peopleare more natural to perceive speech information than screen-displayed messages and can alsoreact more quickly and correctly to speech cues,and if speech input modality is adopted ap-propriately in multiple- task situation,operator's workload can be reduced and system per-formance can be improved.In conclusion,speech technology is feasible for enhancing man-machine system performance.However,adequate considerations should be made in terms ofthe interactions among such factors as human operator,machine,environment and task de-mands.

本研究将汉语语音输入与输出技术应用于计算机模拟的控制任务之中,其目的是考察在人-机系统中利用语音技术来提高操作控制工效的可行性。首先我们以微型计算机为基础,利用语音输入(识别)输出(合成)装置,操纵杆和模数转换器,采用汇编和C 语言混合编程,建立了一个多功能参数可拟式人-机交互控制的实验系统。然后,利用该系统进行双自由度目标控制以及在控制过程中进行数据输入的双重作业实验。根据目标提示和数字输入的不同方式,实验设计了语显键入(SK)、字显键入(CK)、语显语入(SS)和字显语入(CS)等四种作业。实验中,作业安排的次序完全随机化,避免了因实验设计造成的“非对称转移”效应。实验结果表明,作业SK 的平均操作反应时间(ORT)比作业CK 快6.3%, 作业SS 平均ORT 比作业CS 快5.8%;四种作业中,作业SS 的平均ORT 最短,个体差异最小,作业CK 的平均ORT 最长,个体差异最大。从对被试者总体操作反应错误的统计来看,被试者对语音显示的操作反应错误率(ORER)(2.75%)比对屏幕字符显示的ORER(5.38%)要低得多,但被试者用语音输入时的ORER(2.98%)却高于键盘输入方式(其ORER 为...

本研究将汉语语音输入与输出技术应用于计算机模拟的控制任务之中,其目的是考察在人-机系统中利用语音技术来提高操作控制工效的可行性。首先我们以微型计算机为基础,利用语音输入(识别)输出(合成)装置,操纵杆和模数转换器,采用汇编和C 语言混合编程,建立了一个多功能参数可拟式人-机交互控制的实验系统。然后,利用该系统进行双自由度目标控制以及在控制过程中进行数据输入的双重作业实验。根据目标提示和数字输入的不同方式,实验设计了语显键入(SK)、字显键入(CK)、语显语入(SS)和字显语入(CS)等四种作业。实验中,作业安排的次序完全随机化,避免了因实验设计造成的“非对称转移”效应。实验结果表明,作业SK 的平均操作反应时间(ORT)比作业CK 快6.3%, 作业SS 平均ORT 比作业CS 快5.8%;四种作业中,作业SS 的平均ORT 最短,个体差异最小,作业CK 的平均ORT 最长,个体差异最大。从对被试者总体操作反应错误的统计来看,被试者对语音显示的操作反应错误率(ORER)(2.75%)比对屏幕字符显示的ORER(5.38%)要低得多,但被试者用语音输入时的ORER(2.98%)却高于键盘输入方式(其ORER 为1.13%),此系由语音识别系统的性能限制及个别被试者语音变异性所致。这些结果说明,人们对语音显示信息的反应比字符屏显的反应明显的快且对接收语音命令更习惯自然、反应更准确,在多重作业中合适地加入语音输入方式能减轻操作员的工作负荷,从而比单纯的键控方式能获得更高的效率。我们认为,语音技术在人-机系统中的应用潜力很大,对于操作员视觉、手控和认知负荷较重的任务,合理地利用语音输入输出方式替代部分视觉和手控的功能,可望减轻操作人员的工作负荷,提高操作控制工效。同时,我们还需认识到语音装置应用不利的一面,那就是语音信号本身占时长,尤其对语音识别而言,其识别时间可能比手控方式长,出错率较高,抗噪性较差。因此,其应用场合必须谨慎选择,比如,对于需快速反应或应急的场合,便不宜使用语音装置。总之,语音输入输出技术提供了视觉和手控的一种替代方式,符合人的特点,将其用于提高人-机系统工效是可行的,但尚需对人、机、环境以及作业因素的相互作用作充分考虑。

 
<< 更多相关文摘    
图标索引 相关查询

 


 
CNKI小工具
在英文学术搜索中查有关识别输出的内容
在知识搜索中查有关识别输出的内容
在数字搜索中查有关识别输出的内容
在概念知识元中查有关识别输出的内容
在学术趋势中查有关识别输出的内容
 
 

CNKI主页设CNKI翻译助手为主页 | 收藏CNKI翻译助手 | 广告服务 | 英文学术搜索
版权图标  2008 CNKI-中国知网
京ICP证040431号 互联网出版许可证 新出网证(京)字008号
北京市公安局海淀分局 备案号:110 1081725
版权图标 2008中国知网(cnki) 中国学术期刊(光盘版)电子杂志社