-
公开(公告)号：KR1020090112882A
公开(公告)日：2009-10-29
申请号：KR1020080038636
申请日：2008-04-25
Applicant: 목포대학교산학협력단 , (주)이후맥스 , (주)가민정보시스템
Abstract: PURPOSE: A multimedia data offering service using TTS(Text To Speech) and a talking head is provided to apply multimedia contents according to a user purpose. CONSTITUTION: A multimedia data offering service using TTS(Text To Speech) and a talking head comprises the following steps. Desired data is selected from a multimedia medium. It is determined whether to use one or both of TTS(400) or a talking head about the selected data(200). A 3D module talking head(300) is associated with a TTS DB(410) to generate and play a face image of a mouth shape when speaking. A DB expressing an emotion to the TTS is added like an xml tag to increase efficiency of meaning transfer.
Abstract translation: 目的：提供使用TTS（文字到语音）和通话头的多媒体数据提供服务，以根据用户目的应用多媒体内容。 构成：使用TTS（文字到语音）和通话头的多媒体数据提供服务包括以下步骤。 从多媒体媒体中选择所需的数据。 确定是否使用关于所选数据（200）的TTS（400）或谈话头中的一个或两个。 3D模块通话头（300）与TTS DB（410）相关联，以在发言时产生和播放口形的脸部图像。 表示对TTS的情绪的DB被添加为xml标签，以提高意义传递的效率。
-
Patent Agency Ranking
公开(公告)号：KR101081861B1
公开(公告)日：2011-11-09
申请号：KR1020090058578
申请日：2009-06-29
Applicant: 목포대학교산학협력단 , 주식회사세오
Abstract: 본발명은폭행상황검출방법에관한것으로, 보다구체적으로는입력되는프레임영상내의움직이는객체중 사람객체를검출하고, 검출된사람객체들각각의움직임벡터, 상기사람객체들간의위치벡터및 상기프레임영상의음성정보를분석하여폭행상황여부를판단하여, 폭행상황발생시에영상메시지나음성메시지를외부로출력함으로써, 감시자에게폭행상황발생을알려줄 수있는움직이는객체의영상정보에기반한폭행상황검출방법에관한것이다.
-
公开(公告)号：KR1020100115003A
公开(公告)日：2010-10-27
申请号：KR1020090033488
申请日：2009-04-17
Applicant: 목포대학교산학협력단
CPC classification number: G06T13/80 , G06F17/3074 , G06F17/30905 , G06T2207/30201 , G10L19/0018
Abstract: PURPOSE: A face animation output method of a text data base and a system thereof for synchronizing a mouth shape image of the face animation are provided to increase the information transmission by outputting a face animation image. CONSTITUTION: A text data is text data of a power point on program. A text to speech engine(100) receives text data from the power point program. An FAP converter(600) converts a response message data into the parametric value of MPEG-4. A sink module(200) transmits an audio file to the audio output device.
Abstract translation: 目的：提供一种用于同步面部动画的嘴形图像的文本数据库及其系统的面部动画输出方法，以通过输出面部动画图像来增加信息传输。 构成：文本数据是程序上电源点的文本数据。 文本到语音引擎（100）从电源点程序接收文本数据。 FAP转换器（600）将响应消息数据转换为MPEG-4的参数值。 宿模块（200）将音频文件发送到音频输出设备。
-
公开(公告)号：KR101039668B1
公开(公告)日：2011-06-08
申请号：KR1020090033488
申请日：2009-04-17
Applicant: 목포대학교산학협력단
Abstract: 본 발명은 텍스트데이터 기반의 얼굴 애니메이션 출력방법 및 그 시스템에 관한 것으로, 보다 구체적으로는 텍스트데이터의 음소들이 음성으로 출력되기 전에 상기 음소들에 대응하는 얼굴 애니메이션을 렌더링하여 영상으로 출력해줌으로써 상기 음소들의 음성 출력과 상기 얼굴 애니메이션의 입 모양 영상이 서로 일치되도록 동기화할 수 있는 텍스트데이터 기반의 얼굴 애니메이션 출력방법 및 그 시스템에 관한 것이다.
토킹 헤드, 텍스트데이터, 렌더링, 음소, Text-to-speech-
公开(公告)号：KR1020110001172A
公开(公告)日：2011-01-06
申请号：KR1020090058578
申请日：2009-06-29
Applicant: 목포대학교산학협력단 , 주식회사세오
CPC classification number: G06K9/00362 , G06K9/00778 , G06K9/4642 , G06T5/20 , G06T7/194 , G06T7/20 , G10L15/02 , G10L2015/027
Abstract: PURPOSE: A violence detection method by analyzing an motion image of moving peoples is provided to accurately detect a violence situation by judging the voice information of a frame image and a variation vector within person objects within the frame image. CONSTITUTION: A background detection module detects person objects among the objects moving in a present frame image(S1000). If a center coordinate of head tracking boxes is placed in a predetermined distance, an object tracking module selects the present frame image as a violence candidate image(S3000). The object tracking module induces a change vector through acceleration information of a second head moving vector and the angle difference between head moving vectors(S4000b). If the value of the first variation vector is changed into the value within a predetermined range, the object tracking module judges the current situation as a violent state(S5000).
Abstract translation: 目的：提供一种通过分析移动人物的运动图像的暴力检测方法，通过判断帧图像内的人物内的帧图像的声音信息和变化矢量来准确地检测暴力情况。 构成：背景检测模块检测在当前帧图像中移动的对象中的人物（S1000）。 如果头跟踪盒的中心坐标被放置在预定距离内，则对象跟踪模块选择当前帧图像作为暴力候选图像（S3000）。 物体跟踪模块通过第二头部移动矢量的加速度信息和头部移动矢量之间的角度差（S4000b）来引起变化矢量。 如果第一变化矢量的值变为预定范围内的值，则对象跟踪模块将当前情况判断为暴力状态（S5000）。
-
公开(公告)号：KR1020090112218A
公开(公告)日：2009-10-28
申请号：KR1020080037981
申请日：2008-04-24
Applicant: 목포대학교산학협력단
CPC classification number: G06K9/00221 , G06T11/00 , G06T2219/004
Abstract: PURPOSE: A labeling technique for realizing face detection hardware is provided to be utilized in various intelligent system fields. CONSTITUTION: A labeling technique for realizing face detection hardware comprises a step of storing the current re-sampled value(DATA A) and pixel values DATA B and the DATA C in a line memory. If the values are DATA B and DATA A and the DATA A and DATA C, the DATA A adds 1 to the label addresses attached until the present and adds 1 to the label value. If the values are DATA B and DATA A and the DATA A is identical to DATA C, the address of DATA C is attached to DATA A.
Abstract translation: 目的：提供用于实现面部检测硬件的标签技术，用于各种智能系统领域。 构成：用于实现面部检测硬件的标记技术包括将当前重新采样值（DATA A）和像素值DATA B和DATA C存储在行存储器中的步骤。 如果值为DATA B和DATA A，DATA A和DATA C，则DATA A将附加的标签地址加1，直到当前的标签地址，并将1加到标签值。 如果值为DATA B和DATA A，DATA A与DATA C相同，则DATA C的地址连接到DATA A.
-
-
-
-
-
Search Results
6
records
(took 0.017 seconds)
