CN103117058B - Smart TV platform, multi-voice switching system and method based on engine - Google Patents

Smart TV platform, multi-voice switching system and method based on engine Download PDF

Info

Publication number
CN103117058B
CN103117058B CN201210558320.XA CN201210558320A CN103117058B CN 103117058 B CN103117058 B CN 103117058B CN 201210558320 A CN201210558320 A CN 201210558320A CN 103117058 B CN103117058 B CN 103117058B
Authority
CN
China
Prior art keywords
speech
module
engine
voice
speech engine
Prior art date
Application number
CN201210558320.XA
Other languages
Chinese (zh)
Other versions
CN103117058A (en
Inventor
陈冠霖
赵波
刘贤洪
杨金峰
毕端
Original Assignee
四川长虹电器股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 四川长虹电器股份有限公司 filed Critical 四川长虹电器股份有限公司
Priority to CN201210558320.XA priority Critical patent/CN103117058B/en
Publication of CN103117058A publication Critical patent/CN103117058A/en
Application granted granted Critical
Publication of CN103117058B publication Critical patent/CN103117058B/en

Links

Abstract

本发明涉及智能电视软件平台,其公开了一种基于智能电视平台的多语音引擎切换方法,实现自动查找当前识别效率最高的语音引擎并进行切换,提升用户的语音交互体验。 The present invention relates to a smart TV software platform, which discloses a multi-speech engine switching method based smart TV platform, automatically find the current speech engine to identify the most efficient and switching, enhance the user's voice interactive experience. 该方法可以概括为:当用户运行语音应用程序使用语音识别功能时,语音引擎选择模块通过语音应用接口获取采集到的语音数据,然后将语音数据发送给每一个语音引擎模块,记录并比较各个语音引擎模块返回识别结果的响应时间,选择响应时间最短的语音引擎模块进行切换。 The method can be summarized as follows: when the respective voice users running the voice application uses the voice recognition function, a speech engine selection module acquired by the voice application interface collected voice data, and then transmits the voice data to each speech engine module, recorded and compared engine module returns the response time of the identification result, selecting the shortest response time to switch the speech engine module. 此外,本发明还公开了相应的切换系统,适用于在智能电视中实现快速语音识别功能。 Further, the present invention also discloses a corresponding switching system for speech recognition in rapid smart TV.

Description

基于智能电视平台的多语音引擎切换系统及方法 Smart TV platform, multi-voice switching system and method based on engine

技术领域 FIELD

[0001] 本发明涉及智能电视软件平台,具体的说,是涉及一种基于智能电视平台的多语音弓I擎切换系统及方法。 [0001] The present invention relates to a smart TV software platform, particularly, relates to a multi-based speech engine Smart TV platform I bow switching system and method.

背景技术 Background technique

[0002] 随着电视终端智能化、网络化的发展,智能电视可获取的内容得到了极大的丰富,功能也更加的多元化,电视的操控随之变得更加频繁和复杂。 [0002] As the television intelligent terminals, network development, Smart TV content available has been greatly enriched, features and more diversified, manipulation of television are becoming more frequent and complex. 语音识别技术在智能电视上的应用大大简化了用户的操作过程,用户体验得到极大提高。 Speech recognition technology in the smart TV greatly simplifies the user's operation, the user experience has been greatly improved. 由于语音识别需要占用巨大的系统资源,智能电视目前一般都通过网络连接云端服务器来实现语音识别功能; Since voice recognition requires a huge system resources, the smart TV is generally to voice recognition and cloud server through the network;

[0003] 在服务器中用于实现语音识别功能的语音识别引擎由语音检测模块、特征提取模块和识别搜索模块组成;其中,语音检测模块的功能是进行语音信号的检测和与处理,电视将采集到的原始语音数据送入到该模块,语音信号数据需要在语音检测模块里转换成标准的数据格式(比如:8K,16bit);同时,利用高效的信号检测算法,判断出语音的起始点和终止点;特征提取模块收到检测后的语音数据流,从中提取得到语音信号的特征矢量流。 [0003] for realizing the voice recognition function in a server speech recognition engine by the speech detection extraction module, a feature module and a search module identification; wherein the function is voice detection module detecting and processing speech signals, collected TV fed to the original voice data to the module, the data need to be converted voice signal in the voice detection module in a standard data format (example: 8K, 16bit); at the same time, using an efficient signal detection algorithm, and determines the starting point of the speech termination point; wherein the feature extraction module receives the voice data stream is detected, the speech signal is extracted to obtain vector flow. 语音特征是利用数字信号处理技术,从语音信号中提取最反应其本质属性的信息。 Speech features using digital signal processing technology, most information extraction reactor whose essential attributes from the speech signal. 在这个模块中,需要对语音信号进行预加重、分帧、加窗、品与变换、倒谱变换、差分等处理,最终得到数十维左右的特征矢量;识别搜索模块将收到的未知语音信号特征与引擎内的声学模型库、词典/字典和识别语法信息进行匹配,得到最适合未知语音特征的词序列。 In this module, the need for pre-emphasis of the speech signal, framing, windowing, and the transform products, cepstrum transformation, processing such as differential, the final feature vector obtained several tens dimension; unknown speech recognition module searches the received an acoustic model repository with the signal characteristics of the engine, dictionary / grammar dictionary and identification information matches, to obtain the most suitable for the unknown word sequence of speech features. 这个过程可以简单描述如下:通过检索词典/字典,可以将句子由词序列分解成音素的序列。 This process can be described as follows: by searching the dictionary / dictionary, a sentence can be decomposed into a sequence of phonemes of a word sequence. 这种音素的序列与声学模型相结合,就得到更反映其本质属性的声学模型单元序列信息。 Such acoustic models and phoneme sequence combination unit acoustic model is obtained sequence information also reflects its essential attributes. 然后,将原始语音的特征矢量与所有可能的句子候选的声学模型单元序列的信息相互匹配,计算得到其匹配概率,从中挑选出具有最大后验概率的声学模型单元序列。 Then, the original speech feature vector matches the acoustic model information unit sequence of all possible candidate sentences each other, which matched the calculated probability, selected from the acoustic model unit sequence having the maximum posterior probability. 通过该单元序列,可以得到与之对应的词序列,这就是引擎输出给电视的文字序列。 By this means a sequence, it is possible to obtain the corresponding word sequence, which is the sequence of words of the engine output to a television.

[0004] 而由于服务器中存在多个语音识别引擎,如果单一的使用某一个固定引擎进行语音识别,不利于智能电视语音识别效率的提升,造成用户语音交互体验不好;因此,如何在多个语音识别引擎之间查找当前最有效率的语音识别引擎并进行切换是语音交互应用中一个亟待解决的问题。 [0004] And because the plurality of speech recognition engines exist on the server, if a single fixed using a speech recognition engine, intelligent speech recognition is not conducive to enhancing the efficiency of the TV, causing poor voice user interaction experience; therefore, how in multiple Find the most efficient current speech recognition engine between the speech recognition engine and switch voice interaction application is a serious problem.

发明内容 SUMMARY

[0005] 本发明所要解决的技术问题是:提出一种基于智能电视平台的多语音引擎切换系统及方法,实现自动查找当前识别效率最高的语音引擎并进行切换,提升用户的语音交互体验。 [0005] The present invention solves the technical problem: A new Smart TV platform based on multi-engine voice switching system and method for automatically find the current engine to identify the most efficient voice and switching, enhance the user's voice interactive experience.

[0006] 本发明解决上述技术问题采用的方案是:基于智能电视平台的多语音引擎切换系统,包括:语音引擎选择模块及至少两个语音引擎模块;所有的语音引擎模块由统一的语音引擎接口进行封装,并通过语音引擎接口连接语音引擎选择模块;所述语音引擎选择模块通过语音应用接口与语音应用程序相连。 Scheme employed [0006] The present invention is to solve the above technical problem: switching system based on intelligent TV platform multi-speech engine, comprising: a speech engine selection module, and at least two speech engine module; all speech engine module by a uniform interface to the speech engine encapsulating, by the speech engine module interface connected to the voice selection engine; engine selects the voice application module is connected through a voice interface to a voice application.

[0007] 进一步,所述语音引擎模块用于从语音引擎接口获取语音引擎选择模块传送的语音数据,并对语音数据进行识别,然后向语音引擎选择模块返回识别结果;所述语音引擎选择模块用于在语音应用程序使用语音识别功能时,通过语音应用接口获取采集到的语音数据,将语音数据通过语音引擎接口发送给每一个语音引擎模块,并接收所有语音引擎模块返回的识别结果,记录各个语音引擎模块返回识别结果的响应时间并进行对比,选择响应时间最短的语音引擎模块进行切换,使得语音应用程序可以调用到识别效率最高的语音引擎模块。 [0007] Further, the speech engine module is an interface for acquiring speech engine speech data transmitted from the selection module speech engine recognition and voice data, then returns the result to the speech recognition engine selection module; engine selects the voice module at the time of the voice applications using voice recognition function, the interface acquired by the voice application acquired the voice data, the voice data is sent to each of the speech engine module via speech engine interface, and receives a recognition result of all the speech engine module returns recorded each the speech engine module returns the recognition result and compare the response time, selecting the shortest response time to switch the speech engine module, so that the speech recognition application can call to the most efficient speech engine module.

[0008] 进一步,所述选择响应时间最短的语音引擎模块进行切换是指:语音引擎选择模块通过语音引擎接口连接到响应时间最短的语音引擎模块,同时断开与其它语音引擎模块的连接。 [0008] Further, the selection of the shortest response time to switch the speech engine module means: speech engine selection module connected through an interface to the speech engine shortest response time of the speech engine module, and disconnect the other speech engine module.

[0009] 此外,本发明还提出了一种相应的基于智能电视平台的多语音引擎切换方法,包括: [0009] Further, the present invention also proposes a corresponding plurality of TV platform based on intelligent speech engine switching method, comprising:

[0010] a.当用户运行语音应用程序使用语音识别功能时,语音引擎选择模块通过语音应用接口获取采集到的语音数据; . [0010] a user running a voice application using speech recognition, the speech engine selection module acquires the voice acquired by the voice data when the application interface;

[0011 ] b.语音引擎选择模块将语音数据通过语音引擎接口发送给每一个语音引擎模块; . [0011] b speech engine selection module transmits the voice data through the voice engine interface to each speech engine module;

[0012] c.各个语音引擎模块对语音数据进行识别,然后向语音引擎选择模块返回识别结果; . [0012] c engine module respective voice recognition voice data, a selection module and returns the recognition result to the speech engine;

[0013] d.语音引擎选择模块记录各个语音引擎模块返回识别结果的响应时间并进行对比,选择响应时间最短的语音引擎模块进行切换。 [0013] d. Recorded speech engine selection module respective voice recognition engine module returns the response times and the results are compared to select the shortest response time to switch the speech engine module.

[0014] 进一步,步骤d中,所述选择响应时间最短的语音引擎模块进行切换是指:语音引擎选择模块通过语音引擎接口连接到响应时间最短的语音引擎模块,同时断开与其它语音引擎模块的连接。 [0014] Further, in step d, the selection of the shortest response time to switch the speech engine module refers to: a voice selection module connected to the engine in response to engine shortest speech voice engine interface module, and disconnect the other speech engine module Connection.

[0015] 本发明的有益效果是:通过对各个语音引擎模块返回识别结果的响应时间(即识别速度)进行对比,选择响应时间最短的语音引擎模块进行切换,使得语音应用程序可以调用到识别效率最高的语音引擎模块进行语音识别,从而提升了语音识别的整体识别效率;并且,由于语音应用程序与语音引擎选择模块之间的连接载体(语音应用接口)保持不变,当语音引擎模块发生切换时,语音应用程序无需关注具体是哪一个语音引擎模块发生切换,从而保证了语音识别的稳定性和延续性。 [0015] Advantageous effects of the present invention are: comparing by returning a recognition result for each speech engine module response time (i.e., recognition speed), select the shortest response time of the speech engine module is switched, so that the speech application can call identification efficiency highest speech engine module for speech recognition, thereby improving the overall efficiency of the speech recognition of the recognition; and, remains unchanged since the connection carrier (speech application Interface) between the voice application module and the speech engine selection, handoff occurs when the speech engine module , the voice application without concern for a particular speech engine module which switching occurs, thus ensuring the stability and continuity of the speech recognition.

附图说明 BRIEF DESCRIPTION

[0016]图1为本发明中基于智能电视平台的多语音引擎切换系统实现构架图; [0016] Figure 1 is a switching system architecture implemented in FIG invention based on intelligent TV platform multiple speech engine;

[0017]图2为本发明中的基于智能电视平台的多语音引擎切换方法的流程图。 [0017] The flowchart of a handover method based on intelligent multi-speech engine TV platform 2 of the present invention in FIG.

具体实施方式 Detailed ways

[0018] 本发明的实现原理是:由于系统中各个语音引擎模块的性能差异,这些模块对语音数据的处理就有快有慢,因此,我们可以通过设置一个语音引擎选择模块来对各个语音引擎模块处理语音数据的响应时间进行记录和比较,从而找出处理时间最短、响应最快的语音引擎模块,然后切换至该语音引擎模块的连接即可,而语音引擎选择模块的引入由于其与语音应用程序之间的应用接口始终未发生改变,因此,同时还能解决系统的稳定性问题。 [0018] The principle of the present invention are: the performance differences due to the respective system speech engine module, these modules have processed voice data is fast or slow, and therefore, we can set a speech engine for each selection module to speech engine and a comparison module for recording the voice data processing response time, and to find the shortest processing time, the fastest response to the speech engine module, and then switch to a connection to the speech engine module, the speech engine is introduced due to the selection module and the speech the application interface between applications always not changed, and therefore, while also solve the problem of the stability of the system.

[0019] 参见图1,本发明中基于智能电视平台的多语音引擎切换系统包括语音引擎选择模块及多个语音引擎模块;所有的语音引擎模块由统一的语音引擎接口进行封装,并通过语音引擎接口连接语音引擎选择模块;所述语音引擎选择模块通过语音应用接口与语音应用程序相连。 [0019] Referring to Figure 1, the present invention includes a selection module and a plurality of speech engine speech engine speech engine module multiple switching platform based on intelligent television system; all speech engine module is encapsulated by a uniform interface to the speech engine, and by the speech engine speech engine interface selection module; a selection module connected by a speech engine speech application interfaces with the voice application.

[0020] 其中,所述语音引擎模块用于从语音引擎接口获取语音引擎选择模块传送的语音数据,并对语音数据进行识别,然后向语音引擎选择模块返回识别结果;所述语音引擎选择模块用于在语音应用程序使用语音识别功能时,通过语音应用接口获取采集到的语音数据,将语音数据通过语音引擎接口发送给每一个语音引擎模块,并接收所有语音引擎模块返回的识别结果,记录各个语音引擎模块返回识别结果的响应时间并进行对比,选择响应时间最短的语音引擎模块进行切换,使得语音应用程序可以调用到识别效率最高的语音引擎模块。 [0020] wherein the speech engine module is an interface for acquiring speech engine speech data transmitted from the selection module speech engine recognition and voice data, then returns the result to the speech recognition engine selection module; engine selects the voice module at the time of the voice applications using voice recognition function, the interface acquired by the voice application acquired the voice data, the voice data is sent to each of the speech engine module via speech engine interface, and receives a recognition result of all the speech engine module returns recorded each the speech engine module returns the recognition result and compare the response time, selecting the shortest response time to switch the speech engine module, so that the speech recognition application can call to the most efficient speech engine module.

[0021] 图2给出了切换方法的相应流程,其包括以下实现步骤: [0021] Figure 2 shows the corresponding flow switching method, comprising the steps implemented:

[0022] a.当用户运行语音应用程序使用语音识别功能时,语音引擎选择模块通过语音应用接口获取采集到的语音数据;该语音数据来源于智能电视的语音采集设备采集到得音源信号; [0022] a voice when the user runs an application using speech recognition, the speech engine selection module acquires speech data acquired by the voice application interface; The voice data from the voice collecting devices of intelligent television audio signal to obtain;

[0023] b.语音引擎选择模块将语音数据通过语音引擎接口发送给每一个语音引擎模块;由于采用了统一的语音引擎接口进行封装,每一个语音引擎模块都能同时收到同样的语音数据; . [0023] b speech engine selection module transmits the voice data through the voice engine interface to each speech engine module; As a result of the speech engine unified interface package, each of the speech engine module can simultaneously receive the same voice data;

[0024] c.各个语音引擎模块对语音数据进行识别,然后向语音引擎选择模块返回识别结果; . [0024] c engine module respective voice recognition voice data, a selection module and returns the recognition result to the speech engine;

[0025] d.语音引擎选择模块记录各个语音引擎模块返回识别结果的响应时间并进行对比,选择响应时间最短的语音引擎模块进行切换:语音引擎选择模块通过语音引擎接口连接到响应时间最短的语音引擎模块,同时断开与其它语音引擎模块的连接。 . [0025] d speech engine selection module records each speech engine module returns the recognition result in response time and compare, select shortest response time of the speech engine module to switch: the shortest speech speech engine selection module connected to the response time by the speech engine interface engine module, and disconnect the other speech engine module. 此后,语音应用程序都可以通过调用该响应时间最短的语音引擎模块实现快速的语音识别,提升用户的语音交互体验。 Then, the speech application can achieve rapid speech recognition by calling the response time of the shortest speech engine module, to enhance the user's voice interactive experience.

Claims (2)

1.基于智能电视平台的多语音引擎切换系统,其特征在于,包括:语音引擎选择模块及至少两个语音引擎模块;所有的语音引擎模块由统一的语音引擎接口进行封装,并通过语音引擎接口连接语音引擎选择模块;所述语音引擎选择模块通过语音应用接口与语音应用程序相连; 所述语音引擎模块用于从语音引擎接口获取语音引擎选择模块传送的语音数据,并对语音数据进行识别,然后向语音引擎选择模块返回识别结果;所述语音引擎选择模块用于在语音应用程序使用语音识别功能时,通过语音应用接口获取采集到的语音数据,将语音数据通过语音引擎接口发送给每一个语音引擎模块,并接收所有语音引擎模块返回的识别结果,记录各个语音引擎模块返回识别结果的响应时间并进行对比,选择响应时间最短的语音引擎模块进行切换,使得语音应用程序可 1. Multi-speech engine switching platform based on intelligent TV system, characterized by comprising: a speech engine selection module, and at least two speech engine module; all speech engine module is encapsulated by a uniform interface to the speech engine, and an engine interface by voice selection module connected speech engine; engine selects the voice module is connected by voice and speech application program interface to the application; the speech engine module configured to acquire the voice data transmitted by the speech engine module selected from the speech engine interface, and the speech recognition data, then returns the result to the speech recognition engine selection module; means for selecting said speech engine when the voice application uses a voice recognition function, a voice acquired through the application interface to the collected voice data, the voice data is sent to each engine through a voice interface to speech engine module, and receives a recognition result returned by all the speech engine module, the speech engine module returns record individual identification result and compare response time, selecting the shortest response time to switch the speech engine module, so that the speech application can 以调用到识别效率最高的语音引擎模块; 所述选择响应时间最短的语音引擎模块进行切换是指:语音引擎选择模块通过语音引擎接口连接到响应时间最短的语音引擎模块,同时断开与其它语音引擎模块的连接。 To identify the most efficient calls to the speech engine module; selecting the shortest response time to switch the speech engine module refers to: a voice selection module connected to the engine in response to engine shortest speech voice engine interface module, and disconnect the other speech engine module is connected.
2.基于智能电视平台的多语音引擎切换方法,应用在如权利要求1所述的系统中,其特征在于,包括: a.当用户运行语音应用程序使用语音识别功能时,语音引擎选择模块通过语音应用接口获取采集到的语音数据; b.语音引擎选择模块将语音数据通过语音引擎接口发送给每一个语音引擎模块; c.各个语音引擎模块对语音数据进行识别,然后向语音引擎选择模块返回识别结果; d.语音引擎选择模块记录各个语音引擎模块返回识别结果的响应时间并进行对比,选择响应时间最短的语音引擎模块进行切换; 步骤d中,所述选择响应时间最短的语音引擎模块进行切换是指:语音引擎选择模块通过语音引擎接口连接到响应时间最短的语音引擎模块,同时断开与其它语音引擎模块的连接。 2. The handover method based on intelligent TV platform multi-speech engine, according to the application in the system as claimed in claim 1, characterized by comprising:. A voice when the user runs applications use speech recognition, the speech engine module selection speech application Interface acquired collected voice data;. b speech engine selection module transmits the voice data through the voice engine interface to each speech engine module;. c each speech engine module for speech data recognition, and then returns to the speech engine selection module recognition result;. d speech engine selection module records each voice recognition engine module returns the response times and the results are compared to select the shortest response time to switch the speech engine module; step d, the selection of the shortest response time for the speech engine module switching means: speech engine selection module connected through an interface to the speech engine shortest response time of the speech engine module, and disconnect the other speech engine module.
CN201210558320.XA 2012-12-20 2012-12-20 Smart TV platform, multi-voice switching system and method based on engine CN103117058B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210558320.XA CN103117058B (en) 2012-12-20 2012-12-20 Smart TV platform, multi-voice switching system and method based on engine

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210558320.XA CN103117058B (en) 2012-12-20 2012-12-20 Smart TV platform, multi-voice switching system and method based on engine

Publications (2)

Publication Number Publication Date
CN103117058A CN103117058A (en) 2013-05-22
CN103117058B true CN103117058B (en) 2015-12-09

Family

ID=48415416

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210558320.XA CN103117058B (en) 2012-12-20 2012-12-20 Smart TV platform, multi-voice switching system and method based on engine

Country Status (1)

Country Link
CN (1) CN103117058B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103336687B (en) * 2013-06-17 2016-09-14 深圳市金立通信设备有限公司 A method of switching a terminal and an application interface
CN103714814A (en) * 2013-12-11 2014-04-09 四川长虹电器股份有限公司 Voice introducing method of voice recognition engine
CN104795069A (en) * 2014-01-21 2015-07-22 腾讯科技(深圳)有限公司 Speech recognition method and server
CN107018228A (en) * 2016-01-28 2017-08-04 中兴通讯股份有限公司 Voice control system, voice processing method and terminal equipment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1323435A (en) * 1998-10-02 2001-11-21 国际商业机器公司 System and method for providing network coordinated conversational services
CN1429019A (en) * 2001-12-18 2003-07-09 松下电器产业株式会社 TV set with sound discrimination function and its control method
CN1633679A (en) * 2001-12-29 2005-06-29 摩托罗拉公司 Method and apparatus for multi-level distributed speech recognition
CN1723487A (en) * 2002-12-13 2006-01-18 摩托罗拉公司 Method and apparatus for selective speech recognition

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6480819B1 (en) * 1999-02-25 2002-11-12 Matsushita Electric Industrial Co., Ltd. Automatic search of audio channels by matching viewer-spoken words against closed-caption/audio content for interactive television

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1323435A (en) * 1998-10-02 2001-11-21 国际商业机器公司 System and method for providing network coordinated conversational services
CN1429019A (en) * 2001-12-18 2003-07-09 松下电器产业株式会社 TV set with sound discrimination function and its control method
CN1633679A (en) * 2001-12-29 2005-06-29 摩托罗拉公司 Method and apparatus for multi-level distributed speech recognition
CN1723487A (en) * 2002-12-13 2006-01-18 摩托罗拉公司 Method and apparatus for selective speech recognition

Also Published As

Publication number Publication date
CN103117058A (en) 2013-05-22

Similar Documents

Publication Publication Date Title
KR101798828B1 (en) System and method for hybrid processing in a natural language voice services environment
US8775174B2 (en) Method for indexing multimedia information
JP4987203B2 (en) Distributed real-time voice recognition device
US20070043868A1 (en) System and method for searching for network-based content in a multi-modal system using spoken keywords
US20090055185A1 (en) Voice chat system, information processing apparatus, speech recognition method, keyword data electrode detection method, and program
EP2045798A1 (en) Keyword extracting device
CN102196207B (en) Method, device and system for controlling television by using voice
US9348906B2 (en) Method and system for performing an audio information collection and query
US10224024B1 (en) Background audio identification for speech disambiguation
US8015005B2 (en) Method and apparatus for voice searching for stored content using uniterm discovery
CN102549653B (en) Speech translation system, first terminal device, speech recognition server device, translation server device, and speech synthesis server device
CN102799605B (en) An advertisement broadcast monitoring system and method
CN102884569A (en) Integration of embedded and network speech recognizers
US8938393B2 (en) Extended videolens media engine for audio recognition
US8862615B1 (en) Systems and methods for providing information discovery and retrieval
US20080162125A1 (en) Method and apparatus for language independent voice indexing and searching
CN101686286A (en) Realizing method and system of automatic business process with full voice service
CN103236259B (en) Speech recognition processing and a feedback system, voice response method
US9479911B2 (en) Method and system for supporting a translation-based communication service and terminal supporting the service
CN102332262B (en) Method for intelligently identifying songs based on audio features
CN102740014A (en) Voice controlled television, television system and method for controlling television through voice
KR20090130028A (en) Method and apparatus for distributed voice searching
CN101867632A (en) Mobile phone speech instant translation system and method
CN103165131A (en) Voice processing system and voice processing method
JP2003515778A (en) Speech recognition method and apparatus according to a separate language model

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
C14 Grant of patent or utility model