会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 31. 发明公开
    • 디지털 방송 수신기 및 그 제어방법
    • 数字广播接收机及其控制方法
    • KR1020080067443A
    • 2008-07-21
    • KR1020070004734
    • 2007-01-16
    • 삼성전자주식회사
    • 김정수김형남이동윤전재욱
    • H04N21/435H04N21/431
    • H04H60/25H04H60/72
    • A digital broadcasting receiver and a control method thereof are provided to update EPG(Electronic Program Guide) information while outputting a screen saver by using an used tuner and to improve accuracy of the EPG information of each channel, thereby guiding an accurate program to be broadcasted. A method for controlling a digital broadcasting receiver comprises the steps of; counting the time when buttons of a remote control or an input part are not operated(40); judging whether the counted time exceeds a set time(42); judging whether a digital broadcasting signal is received from a tuner(44); outputting a screen saver image stored in a memory through an OSD(On Screen Display) synthesis part to a DTV(46); storing EPG(Electronic Program Guide) information of all channels separately and replacing existing EPG information with the new EPG information; judging whether buttons of the remote control are operated(50); and stopping the output of the screen saver when the buttons are operated, stopping receiving the digital broadcast, and displaying the screen right before the output of the screen saver.
    • 提供数字广播接收机及其控制方法,通过使用所使用的调谐器输出屏幕保护程序来更新EPG(电子节目指南)信息,并提高每个频道的EPG信息的准确性,从而引导准确的节目广播 。 一种用于控制数字广播接收机的方法,包括以下步骤: 计算遥控器或输入部件的按钮未被操作的时间(40); 判断计时时间是否超过设定时间(42); 判断是否从调谐器(44)接收数字广播信号; 通过OSD(屏幕显示)合成部分将存储在存储器中的屏幕保护图像输出到DTV(46); 分别存储所有频道的EPG(电子节目指南)信息,并用新的EPG信息替换现有的EPG信息; 判断遥控器的按钮是否被操作(50); 当操作按钮停止屏幕保护程序的输出,停止接收数字广播,并在屏幕保护程序的输出之前显示屏幕。
    • 33. 发明公开
    • 반도체 웨이퍼의 결함 검출 방법
    • 用于检查半导体波形缺陷的方法
    • KR1020070069605A
    • 2007-07-03
    • KR1020050131922
    • 2005-12-28
    • 삼성전자주식회사
    • 박평선김정수김봉수김영남
    • H01L21/67
    • H01L22/12G01N21/8851G01N2021/8874
    • A method for detecting defects on a semiconductor wafer is provided to accurately measure the size of a defect depending upon a type of the defect by classifying the defect into particle defect and pit defect. Defects existing on a surface of a wafer are classified, and then a curve representing a size of each defect is prepared. The size of each defect is measured by using the curve. The curve comprises a calibration curve prepared by using a laser scattering volume reflected from the defect. The defect comprises a falling defect dropped from the surface of the wafer and a defect dug on the surface.
    • 提供了一种用于检测半导体晶片上的缺陷的方法,以通过将缺陷分类为颗粒缺陷和凹坑缺陷来精确地测量取决于缺陷类型的缺陷尺寸。 将存在于晶片表面上的缺陷分类,然后准备表示每个缺陷的尺寸的曲线。 通过使用曲线测量每个缺陷的尺寸。 该曲线包括通过使用从缺陷反射的激光散射体积制备的校准曲线。 该缺陷包括从晶片表面掉落的下落缺陷和在表面上挖出的缺陷。
    • 34. 发明授权
    • 다 단계 음성 인식 장치 및 방법
    • 多层语音识别装置及方法
    • KR100695127B1
    • 2007-03-14
    • KR1020040080352
    • 2004-10-08
    • 삼성전자주식회사
    • 이재원조정미황광일이영범김정수
    • G10L15/30
    • G10L15/08G10L15/30G10L15/32
    • 다 단계 음성 인식 장치 및 방법이 개시된다. 이 장치는, 인식하고자 하는 음성의 특징을 이용하여 음성을 자체에서 인식 가능한가를 검사하고, 검사된 결과에 따라 음성을 인식하거나 특징을 전송하는 클라이언트 및 제1 ~ 제N 서버들을 구비하고, 제1 서버는 클라이언트로부터 전송된 특징을 이용하여 음성을 자체에서 인식 가능한가를 검사하고, 검사된 결과에 따라 음성을 인식하거나 특징을 전송하고, 제n(2≤n≤N) 서버는 제n-1 서버로부터 전송된 특징을 이용하여 음성을 자체에서 인식 가능한가를 검사하고, 검사된 결과에 따라 음성을 인식하거나 특징을 전송하는 것을 특징으로 한다. 그러므로, 클라이언트의 사용자로 하여금 음성을 고품질로 인식할 수 있도록 하고 예를 들면 연속 음성 인식을 할 수 있고, 클라이언트와 적어도 하나의 서버간의 음성 인식에 따른 부담을 최적으로 분산시켜 음성 인식 속도를 향상시킬 수 있는 효과를 갖는다.
    • 提供了多层语音识别装置和方法。 该装置包括客户端,检查客户端本身是否可以识别语音,使用要识别的语音的特征,并根据检查结果识别语音或发送语音的特征; 以及第一至第N服务器,其中第一服务器使用从客户端发送的语音的特征来检查第一服务器本身是否能够识别语音,并且根据检查结果识别语音或发送特性,并且其中 第n(2‰n‰N)服务器使用从第(n-1)个服务器发送的语音的特征来检查第n个服务器本身是否可以识别语音,并识别语音或 根据检查结果发送特性。
    • 36. 发明授权
    • 음성 인식 장치 및 방법
    • 用于识别语音的装置和方法
    • KR100664960B1
    • 2007-01-04
    • KR1020050094068
    • 2005-10-06
    • 삼성전자주식회사
    • 정상배김남훈한익상김정수최인정
    • G10L15/14G10L15/28
    • G10L15/142
    • A voice recognition apparatus and a method thereof are provided to calculate a hidden Markov model state representing a characteristic parameter of a predetermined unit of a voice by using single waveform probability distribution to obtain a score and calculate a state having a high score using multi-waveform probability distribution to produce a score. A voice recognition apparatus includes a characteristic vector extractor(330), an expected node extractor(350), a single waveform similarity calculator(315), a multi-waveform similarity calculator(320), and an output unit(340). The characteristic vector extractor splits an input voice signal into unit regions and extracts characteristic vectors corresponding to the unit regions. The expected node extractor extracts a list of second nodes which are expected to move to first nodes corresponding to the characteristic vectors with reference to a predetermined network composed of at least one node. The single waveform similarity calculator substitutes the characteristic vectors for single waveform probability distribution constructing a voice signal corresponding to the second nodes to obtain single waveform similarities between the first nodes and the second nodes of the list. The multi-waveform similarity calculator substitutes the characteristic vectors for multi-waveform probability distribution constructing single waveform probability distribution which produces single waveform similarity included in a predetermined range among the single waveform similarities to produce multi-waveform similarities. The output unit outputs a function execution signal corresponding to multi-waveform probability distribution which produces the highest value among the multi-waveform similarities.
    • 提供一种语音识别装置及其方法,通过使用单波形概率分布来计算表示声音的预定单位的特征参数的隐马尔可夫模型状态,以获得分数并使用多波形计算具有高分数的状态 概率分布产生得分。 语音识别装置包括特征向量提取器(330),预期节点提取器(350),单个波形相似度计算器(315),多波形相似度计算器(320)和输出单元(340)。 特征向量提取器将输入的语音信号分离成单位区域,并提取与单位区域对应的特征向量。 参考由至少一个节点组成的预定网络,期望节点提取器提取期望移动到与特征向量对应的第一节点的第二节点的列表。 单波形相似度计算器代替构成与第二节点对应的语音信号的单波形概率分布的特征向量,以获得列表的第一节点和第二节点之间的单波形相似度。 多波形相似度计算器将构成单波形概率分布的多波形概率分布的特征向量代入单波形相似度中产生预定范围内的单一波形相似度,产生多波形相似度。 输出单元输出对应于在多波形相似度中产生最高值的多波形概率分布的功能执行信号。