-
公开(公告)号：KR1020140005639A
公开(公告)日：2014-01-15
申请号：KR1020120073518
申请日：2012-07-05
Applicant: 삼성전자주식회사
CPC classification number: G10L15/10 , G10L15/183 , G10L15/26
Abstract: Provided are an electronic apparatus and a method for modifying voice recognition errors thereof. The method for modifying voice recognition errors of the electronic apparatus includes recognizing user voice inputted through a voice input part by using a voice model and a language model; displaying the recognized user voice; extracting candidate text for selected text when at least one text among full text displayed is selected; displaying an error correction UI including the candidate text; displaying and changing the selected candidate text as at least one text among candidate text displayed on the error correction UI is selected, and updating the voice model and the language model according to change results. Therefore, a user easily corrects errors due to a voice recognition result. [Reference numerals] (AA) Start; (BB) End; (S610) Inputting user voice; (S620) Recognize the user voice by using a voice model and a language model; (S630) Display the recognized user voice with texts; (S640) Select at least one text?; (S650) Extract candidate text for at least one text; (S660) Display an error correction UI; (S670) Select one among the candidate text; (S680) Change at least one text into the candidate text and display; (S690) Update the voice model and the language model
Abstract translation: 提供一种用于修改其语音识别错误的电子设备和方法。 用于修改电子设备的语音识别错误的方法包括通过使用语音模型和语言模型识别通过语音输入部分输入的用户语音; 显示识别的用户声音; 当选择全文中的至少一个文本时，提取所选文本的候选文本; 显示包括候选文本的纠错UI; 选择在错误校正UI中显示的候补文本中的至少一个文本，并且根据变化结果来更新语音模型和语言模型。 因此，用户容易纠正由于语音识别结果引起的错误。 （附图标记）（AA）开始; （BB）结束; （S610）输入用户声音; （S620）通过使用语音模型和语言模型识别用户语音; （S630）用文字显示识别的用户声音; （S640）至少选一个文本？ （S650）提取至少一个文本的候选文本; （S660）显示纠错用户界面; （S670）在候选文本中选择一个; （S680）将至少一个文本更改为候选文本并显示; （S690）更新语音模型和语言模型
-
公开(公告)号：KR101326262B1
公开(公告)日：2013-11-20
申请号：KR1020070138981
申请日：2007-12-27
Applicant: 삼성전자주식회사
Abstract: 본 발명은 음성인식 단말 및 그 방법에 관한 것으로, 본 발명에 따른 음성인식 단말은 사용자로부터 입력된 음성을 디코딩하여 미확인된 발음열을 생성하는 제1 디코딩부, 생성된 발음열을 서버에 송신하며 송신된 발음열과 관련된 어휘셋을 디코딩한 결과를 서버로부터 수신하는 송수신부 및 서버를 통해 디코딩된 결과 및 생성된 발음열과 관련된 확장 어휘셋을 디코딩하여 미확인된 음성과 가장 유사한 어휘 후보군을 선정하는 제2 디코딩부를 포함하는 것을 특징으로 한다.
음성인식, 발음열, 음성 디코딩, 어휘 디코딩, ASR, DSRAbstract translation: 提供语音识别装置及其方法，通过有效地选择搜索词汇集来准确地识别输入的用户声音。 第一解码单元（100）通过解码由用户输入的语音来产生未识别的发音序列。 收发器（110）将产生的发音序列发送到服务器（20）。 服务器从批量词汇集中提取过滤的词汇集。 第二解码单元（120）选择与未识别的语音最相似的候选词汇组。
-
公开(公告)号：KR1020120080327A
公开(公告)日：2012-07-17
申请号：KR1020110001706
申请日：2011-01-07
Applicant: 삼성전자주식회사
CPC classification number: H04W52/146 , H04W52/241 , H04W68/04 , H04W88/023 , Y02D70/00
Abstract: PURPOSE: An apparatus for efficiently controlling uplink transmission power in a wireless communication system and a method thereof are provided to determine power control value by establishing a target SNR(Signal to Noise Raito) due to a downlink path loss of a terminal. CONSTITUTION: A base station estimates DL(DownLink) path loss(100). The base station determines the target SNR due to the downlink path loss(102). The base station determines power control quantity in order to compensate difference between the target SNR and the estimated SNR(104). The base station transmits power control message including TPC(Transmit Power Control) command value based on the power control value(106).
Abstract translation: 目的：提供一种用于有效控制无线通信系统中的上行发送功率的装置及其方法，用于通过由于终端的下行链路路径损耗建立目标SNR（Signal to Noise Raito）来确定功率控制值。 构成：基站估计DL（DownLink）路径损耗（100）。 基站由于下行链路路径损耗来确定目标SNR（102）。 基站确定功率控制量，以补偿目标SNR与估计SNR之间的差异（104）。 基站基于功率控制值（106）发送包括TPC（发送功率控制）指令值的功率控制消息。
-
公开(公告)号：KR1020120079342A
公开(公告)日：2012-07-12
申请号：KR1020110000576
申请日：2011-01-04
Applicant: 삼성전자주식회사
IPC: G10L15/14
CPC classification number: G10L15/08 , G10L15/065
Abstract: PURPOSE: A sound processing device based on a location and a method thereof are provided to generate a sound model to which environmental features are reflected and process an input voice signal using the generated sound model. CONSTITUTION: A first extracting unit(130) extracts a first sound module matched with a first location set on a voice recognition target area. A second extracting unit(140) extracts a second sound model matched with a second location around the first location. A sound model generating unit(150) generates a third sound model based on at least one of the first and second sound models.
Abstract translation: 目的：提供一种基于位置及其方法的声音处理装置，以产生反映环境特征的声音模型，并使用所产生的声音模型处理输入的语音信号。 构成：第一提取单元（130）提取与在语音识别目标区域上设置的第一位置相匹配的第一声音模块。 第二提取单元（140）提取与围绕第一位置的第二位置匹配的第二声音模型。 声音模型生成单元（150）基于第一和第二声音模型中的至少一个产生第三声音模型。
-
公开(公告)号：KR1020120015374A
公开(公告)日：2012-02-21
申请号：KR1020100072272
申请日：2010-07-27
Applicant: 삼성전자주식회사
CPC classification number: H04M1/7253 , H04M2250/04 , H04W92/18 , H04W8/22
Abstract: PURPOSE: A portable function transition apparatus and a method thereof are provided to enable a peripheral device to control a terminal function in case there is no user feedback. CONSTITUTION: A peripheral terminal generates a notification signal for terminal function execution in case of execution of a terminal function(409, 411). The peripheral terminal transmits a terminal function execution alarm signal to a lower priority device. The peripheral terminal performs corresponding terminal function based on setting information of a terminal function(413). The peripheral device transmits a user feedback absence alarm signal to the lower priority device(415,417).
Abstract translation: 目的：提供便携式功能转换装置及其方法，以便在没有用户反馈的情况下使外围设备能够控制终端功能。 构成：在执行终端功能（409,411）的情况下，外围终端生成用于终端功能执行的通知信号。 外围终端向较低优先级的设备发送终端功能执行报警信号。 外围终端基于终端功能的设定信息进行相应的终端功能（413）。 外围设备向较低优先级的设备发送用户反馈缺失警报信号（415,417）。
-
Patent Agency Ranking
公开(公告)号：KR1020100023375A
公开(公告)日：2010-03-04
申请号：KR1020080082098
申请日：2008-08-21
Applicant: 삼성전자주식회사
IPC: H04B1/40
CPC classification number: H04M1/72552
Abstract: PURPOSE: A message deletion method in a portable terminal and a device thereof are provided to determine message priority to delete a message, thereby conveniently deleting the message. CONSTITUTION: A service message is received(100). The message priority of the service message is calculated(102). If there is no storing space for the service message, pre-stored service messages are displayed based on a predetermined number according to the message priority(104~108). At least one pre-stored service message is selected among the service messages(110). The selected service message is deleted. If there is available space for an incoming message, the incoming message is saved(112,114).
Abstract translation: 目的：提供便携式终端及其装置中的消息删除方法，以确定消息优先级以删除消息，从而方便地删除该消息。 构成：接收到服务消息（100）。 计算服务消息的消息优先级（102）。 如果服务消息没有存储空间，则根据消息优先级（104〜108），基于预定数量显示预存储的服务消息。 在服务消息（110）中选择至少一个预先存储的服务消息。 所选服务消息被删除。 如果传入消息有可用空间，则保存传入消息（112,114）。
-
公开(公告)号：KR1020090122043A
公开(公告)日：2009-11-26
申请号：KR1020080048240
申请日：2008-05-23
Applicant: 삼성전자주식회사
IPC: H04B1/40
CPC classification number: H04N1/32112 , H04N1/21
Abstract: PURPOSE: A method for generating a picture file linked with a schedule management function and a mobile communication terminal thereof are provided to set names of a folder in which a picture file will be stored and a picture file including schedule information of a user during photographing. CONSTITUTION: A camera unit(180) produces image information by capturing a continuous digital signal as a static image. A memory(170) stores a lot of programs necessary for controlling an operation of a mobile communication terminal and a schedule management program of a user. A controller(130) checks whether preregistered schedule information including a picture photographing time exists in the schedule management program of the mobile communication terminal when storing a picture file after photographing. When the pre-registered schedule information exists according to the check result, the controller generates picture file information including the stored schedule information.
Abstract translation: 目的：提供一种用于生成与日程管理功能相关联的图片文件及其移动通信终端的方法，用于设置在其中存储图像文件的文件夹的名称以及包括用户在拍摄期间的时间表信息的图像文件。 构成：相机单元（180）通过捕获作为静态图像的连续数字信号来产生图像信息。 存储器（170）存储用于控制移动通信终端的操作和用户的日程管理程序所需的大量程序。 控制器（130）在拍摄后存储图像文件时，检查包括图像拍摄时间的预先登记的进度信息是否存在于移动通信终端的调度管理程序中。 当根据检查结果存在预先登记的日程信息时，控制器生成包括所存储的日程信息的图片文件信息。
-
公开(公告)号：KR100679051B1
公开(公告)日：2007-02-05
申请号：KR1020050123511
申请日：2005-12-14
Applicant: 삼성전자주식회사
Abstract: A speech recognition apparatus and method using a plurality of confidence score estimation algorithms are provided to set a basis for determining whether a speech recognition result has confidence using a score according to a likelihood ratio and a score using a locus of Gaussian distribution and recognize an inputted speech according to the basis. A speech recognition apparatus includes a first confidence score estimator(150), a second confidence score estimator(160), and a determination unit(190). The first confidence score estimator estimates a first confidence score using a ratio of likelihood of a keyword model with respect to a feature vector for each frame of a speech signal to likelihood of a filler model. The second confidence score estimator compares the locus of Gaussian distribution of the keyword model to the locus of Gaussian distribution of a keyword corresponding to the keyword model to estimate a second confidence score. The determination unit determines whether the estimation result according to the keyword model has confidence based on the point determined by the first and second confidence scores.
Abstract translation: 提供了一种使用多个置信度评分算法的语音识别设备和方法，用于设置用于根据似然比使用分数来确定语音识别结果是否具有置信度的基础以及使用高斯分布轨迹的分数来识别输入 根据基础讲话。 语音识别装置包括第一置信度分数估计器（150），第二置信度分数估计器（160）和确定单元（190）。 第一置信度分数估计器使用关键字模型相对于语音信号的每帧的特征向量与填充模型的似然性的比率来估计第一置信度分数。 第二置信度分数估计器将关键字模型的高斯分布的轨迹与对应于关键字模型的关键字的高斯分布的轨迹进行比较以估计第二置信度分数。 确定单元基于由第一和第二置信度分数确定的分数来确定根据关键字模型的估计结果是否具有置信度。
-
公开(公告)号：KR100557161B1
公开(公告)日：2006-03-03
申请号：KR1019990019791
申请日：1999-05-31
Applicant: 삼성전자주식회사
Inventor： 김남훈
IPC: H04B1/40
Abstract: 가. 청구범위에 기재된 발명이 속한 기술분야
다기능 휴대폰 장치에 관한 기술이다.
나. 발명이 해결하고자 하는 기술적 과제
다기능 휴대폰 장치에 휴대폰 회로를 최대한 활용하며 무전기 겸용 휴대폰 장치를 제공한다.
다. 발명의 해결방법의 요지
본 발명은 무전기 전환키와 송신키를 구비하는 다기능 휴대폰 장치로, 송신할 신호를 무전기 모드와 휴대폰 모드에 공통으로 사용하는 공용 송신부와, 수신되는 신호를 무전기 모드와 휴대폰 모드에 공통으로 사용하는 공용 수신부와, 상기 공용 송신부에서 출력되는 신호를 휴대폰에서 사용하는 무선신호로 변환하여 출력하며, 휴대폰에서 사용하는 무선신호를 수신하여 상기 공용 수신부로 출력하는 휴대폰 송수신부와, 상기 공용 송신부로부터 수신되는 신호를 무전기 대역의 신호로 변환하여 출력하며, 무전기 대역의 신호로 수신되는 신호를 수신하여 상기 공용 수신부로 출력하는 무전기 송수신부와, 상기 공용 송신부로부터 출력되는 신호를 제1스위칭 제어신호에 의해 상기 휴대폰 송수신부 또는 제3스위치의 제1단으로 연결하기 위한 제1스위치와, 상기 휴대폰 송수신부 또는 제3스위치의 제2단을 상기 제1스위칭 제어신호에 의해 상기 공용 수신부로부터 연결하기 위한 제2스위치와, 제2제어신호에 의해 상기 제1단을 무전기의 송신부와 연결하거나 제2단을 무전기의 수신 부로 연결하기 위한 제3스위치와, 상기 공용 송신부와 상기 공용 수신부와 상기 휴대폰 송수신부와 상기 무전기 송수신부로 출력하는 주파수를 발생하는 발진부와, 휴대폰 모드시에 코딩 및 디코딩을 수행하고, 무전기 모드시 음성신호를 처리하는 음성 처리부와, 무전기 모드로 전환시 상기 제1스위칭 신호를 출력하며, 송신키 입력시 상기 제2스위칭 제어신호를 출력하고, 상기 발진부의 주파수 출력을 제어하는 제어부로 구성됨을 특징으로 한다.
라. 발명의 중요한 용도
휴대폰에 사용된다.
휴대폰, 무전기.-
公开(公告)号：KR1020050010482A
公开(公告)日：2005-01-27
申请号：KR1020040044901
申请日：2004-06-17
Applicant: 삼성전자주식회사 , 코닝정밀소재 주식회사
IPC: G02F1/13357
Abstract: PURPOSE: A surface light source device is provided to compensate stress generated on a lower substrate while forming barrier walls on the lower substrate by that generated on the lower substrate while forming a reflection layer thereon, thereby preventing the deformation of the lower substrate formed with the barriers and the reflection layer. CONSTITUTION: A surface light source device includes an upper substrate(3), and a lower substrate(5) corresponding to the upper substrate. A plurality of barrier walls(10) are formed between the upper and lower substrates for forming discharging spaces(11) between the upper and lower substrates. The lower substrate is formed with a reflecting layer(15), and insides of the discharging spaces are respectively formed with fluorescent layers(20). As predetermined electric fields are applied into the discharging spaces from electrodes, plasma discharge is induced into the discharge spaces, generating UV rays. By the UV rays, fluorescent substances is excited from the fluorescent layer, and visible rays are generated from the fluorescent layer. The electrodes surrounds both outer surfaces of the upper or lower substrate.
Abstract translation: 目的：提供一种表面光源装置，用于补偿在下基板上产生的应力，同时在下基板上形成阻挡壁，同时在其上形成反射层，从而防止形成在下基板上的下基板的变形 障碍和反射层。 构成：表面光源装置包括上基板（3）和与上基板对应的下基板（5）。 在上基板和下基板之间形成有多个阻挡壁（10），用于在上基板和下基板之间形成放电空间（11）。 下基板形成有反射层（15），并且排出空间的内部分别形成有荧光层（20）。 当预定电场从电极施加到放电空间中时，等离子体放电被引入放电空间，产生紫外线。 通过紫外线，从荧光层激发荧光物质，从荧光层产生可见光。 电极围绕上或下基板的两个外表面。
-
-
-
-
-
-
-
-
-
Search Results
92
records
(took 0.031 seconds)
