查询词典 recognition marks

Nowadays, speech recognition system in clean conditions with reading input mode can give desirable performance, but environmental noise and speaker variation, such as spontaneousness and accent etc., prevent the recognition engine from being put to real application.

当前在纯净环境下采用朗读方式的语音识别系统识别性能已经达到一个相当高的程度，在这种情况下，语音识别走向实际应用必须解决对噪声环境的鲁棒性。。。

Test results at 20 real intersections show that the ratio of correct stabling siding recognition reaches up to 90%; and the ratios of recognition of traffic lights and divided characters are 85% and 97%, respectively.

通过20个实际交叉口场景测试数据验证，采用所提出的方法，停车线检测正确率达90%，信号灯检测正确率为85%，在信号灯字符正确分割出来的基础上，字符的识别率达97%。

Aiming at the two problems in speech recognition, which are ,(1) The connection between state duration and observed character vector;(2) the dependent problem of frames in speech recognition.

本文从语音识别的实际背景入手，针对两点进行研究：(1)模型状态的驻留时间对于观测特征概率的影响；(2)语音识别中的帧相关问题。

Aiming at the questions of traditional method in pattern recognition of surface defect images of cold steel strip, a pattern recognition method based on wavelet transform and genetic algorithms was proposed.

针对传统的方法在冷轧带钢表面缺陷图像模式识别中存在的问题，提出了基于小波变换和遗传算法的模式识别方法。

Optical Formula Recognition is not only the sign of formula recognition, but also include the extraction of the formula, the record of two dimension structure of the formula and so on.

数学公式的识别不仅只是公式中符号的识别，其还包括了公式抽取，公式二维结构的记录等方面的研究。

Firstly, this article reviews fMRI studies on identity recognition through face and voice processing, including recognition of integrative identity and the subcomponent of identity, such as race, age, sex, and familiarity.

近年来，随着认知神经科学的兴起和发展，采用脑功能磁共振成像(functional Magnetic Resonance Imaging，flVlRI)技术和神经心理学方法(以脑损伤病人为考察对象)对身份识别的研究越来越多，文章将综述这两个领域的研究，并对未来研究方向进行展望。1概述人们在看到他人面孔、听到他人声音的时候，就可提取与此人有关的性别、年龄、熟悉性、职业，甚至个性等个人身份信息，这就是通常所说的身份识别。

Summarized article architecture and pointed out its significance for organizing abstract; Constructed a new method for finding new Chinese word which employs the legend tree technique. The recognition of the elements of document structure and the recognition of multi-level subhead were discussed in detail.

总结了常见的文章体系结构，分析了它对文摘生成的指导意义；提出了一种采用联想树和信息增量技术进行新词识别的方法；对文章结构元素的识别进行了细致地分析，重点构造了多级小标题的识别方法。

Not all nonlinear features are beneficial to the recognition. The superabundant unrelated features may reduce the recognition performance.

然而并非所有非线性特征对识别都有利，过多的不相关特征可能会降低识别性能。

In allusion to the characteristics of the information uncertainty during the sensor management process in fusion target recognition system, we proposed a sensor management method for target recognition based on discrimination.

针对融合目标识别系统传感器管理过程信息不确定性的特点，首次提出了一种基于分辨力的目标识别系统中的传感器管理方法。

At present, from the theoretical research the fingerprint recognition is superior to the recognition tecnology of other living beings.

而其中，根据目前的理论研究来看，指纹识别技术在可操作性上远远优于其它生物识别技术。

The quality of both red and white wine depends very much on the wine maker talent and on the quality of the grapes.

他的红酒和白酒的品质，根据酿造者的能力和葡萄的质量，各有差异。

No, I don't want to get drunk before my next match.

" "不，我不想让我的下一场比赛前喝醉了。