公开(公告)号：KR1020010035786A

公开(公告)日：2001-05-07

申请号：KR1019990042518

申请日：1999-10-02

Applicant: 한국전자통신연구원

Inventor： 이한규 ,  김진웅

IPC: H04N7/01

CPC classification number: H04N19/557 ,  H04N19/44 ,  H04N19/56

Abstract: PURPOSE: An apparatus for indexing image data through joint point-based action information, and an apparatus and a method for searching by using the same are provided to extract joint point in image data and convert it into an expression format including action information for indexing, thereby searching the wanted scene based on the attitude and the action of an object having joint point. CONSTITUTION: Joint point information for extracting joint point data is established(301). Image data to be index is received(302) and it is judged whether the image data is motion picture data(303). If motion picture data, scene change is detected from the motion picture data for generating a shot(304). Joint point data of an object is extracted from the shot(305). If not motion picture data, the object to be obtained joint point data is analyzed in the stop image data for extracting joint point data. The index is performed by analyzing the shot or the stop image(307). The index data is stored(308).

Abstract translation: 目的：提供一种通过基于联动点的动作信息对图像数据进行索引的装置，提供一种使用该图像数据进行搜索的装置和方法，以提取图像数据中的联合点并将其转换成包括用于索引的动作信息的表达格式， 从而基于具有关节点的对象的姿态和动作来搜索想要的场景。 构成：提出联合点数据的联合点信息（301）。 接收要索引的图像数据（302），并且判断图像数据是否是运动图像数据（303）。 如果运动图像数据，则从用于产生拍摄的运动图像数据中检测到场景变化（304）。 从拍摄（305）中提取物体的关节点数据。 如果不是运动图像数据，则在用于提取关节点数据的停止图像数据中分析要获得的关节点数据的对象。 通过分析镜头或停止图像来执行索引（307）。 索引数据被存储（308）。

公开(公告)号：KR101875230B1

公开(公告)日：2018-08-03

申请号：KR1020120012427

申请日：2012-02-07

Applicant: 한국전자통신연구원

Inventor： 이상윤 ,  오천인 ,  이한규 ,  홍진우 ,  정승도 ,  최병욱 ,  이주호 ,  정경부

IPC: H04N21/258

Abstract: 사용자의시청정보를사용하여사용자의선호도를추론하고, 추론한선호도와컨텐츠메타데이터를사용하여맞춤형방송컨텐츠와광고컨텐츠를제공하는장치및 방법이개시된다. 선호도추론장치는단말기로부터사용자의시청정보를수신하는시청정보수신부; 사용자의시청정보를분석하여사용자가관심을가지는컨텐츠를추출하는시청정보분석부; 및추출한컨텐츠를기초로장르와기간에따른사용자의선호도를추론하는선호도추론부를포함할수 있다.

公开(公告)号：KR1020150074858A

公开(公告)日：2015-07-02

申请号：KR1020130163039

申请日：2013-12-24

Applicant: 한국전자통신연구원

Inventor： 오천인 ,  이한규 ,  정영호 ,  홍진우

IPC: G06K9/46

CPC classification number: H04N1/00244 ,  G06K9/00234 ,  G06K9/56 ,  G06T7/254 ,  G06T2207/20081 ,  G06T2207/30201 ,  H04N21/23418 ,  H04N21/4223 ,  H04N21/44008 ,  H04N21/6582

Abstract: 본발명에따른사용자인식전처리장치는수신된시청영상및 기등록된배경영상의비교를통한조명변화여부에기초하여배경영상을갱신하는영상획득제어부, 기등록된배경영상또는갱신된배경영상을 MCT(Modified Census Transform)을통해변환하여 MCT 배경영상을생성하고, 시청영상을 MCT 변환하여 MCT 시청영상을생성하는 MCT 변환부및 MCT 시청영상및 MCT 배경영상을차분한차영상을생성하는차영상처리부를포함한다.

Abstract translation: 根据本发明的用户识别系统包括：图像获取控制单元，通过比较所接收的观看图像和预注册背景图像，基于照明变化来更新背景图像; MCT转换单元，其通过修改的人口普查变换（MCT）转换预注册的背景图像或更新的背景图像，以产生MCT背景图像，并通过MCT转换观看图像以产生MCT观看图像; 以及差分图像处理单元，其利用MCT观看图像和MCT背景图像生成松弛差分图像。

公开(公告)号：KR1020150061716A

公开(公告)日：2015-06-05

申请号：KR1020130145655

申请日：2013-11-27

Applicant: 한국전자통신연구원

Inventor： 박용문 ,  이한규 ,  정영호 ,  홍진우

IPC: G06F19/00

CPC classification number: G06Q30/0242 ,  G06K9/00281

Abstract: 본발명에따른광고효과분석장치는수신된시청자영상으로부터검출된얼굴방향별얼굴코드값에대응하는등록된시청자정보를검색하여시청자를식별하며, 식별된등록된시청자정보및 얼굴방향별얼굴코드값에기초하여식별된시청자의시청행태정보를검출하는시청자인식부및 검출된시청행태정보에기초하여시청자의몰입도를분석하고, 분석된몰입도에기초하여광고효과분석보고서를생성하는광고효과분석부를포함하며, 시청자를촬영하여시청자영상을생성하고, 시청자로부터시청자등록정보를수신하는정보수집부를더 포함할수 있다.

Abstract translation: 根据本发明的用于分析广告效果的设备包括：观看者识别单元，其通过从接收到的观看者图像的面部方向搜索与面部代码相对应的注册观众信息来识别观看者，并且检测观看者的观看行为信息， 基于所识别的注册的观看者信息和面部方向的面部代码值进行识别; 以及广告效果分析单元，其基于检测到的观看行为信息来分析观众的浸入程度，并且基于分析的浸入度生成广告效果分析报告。 本发明还可以包括：通过拍摄观看者生成观看者视频并从观看者接收观看者注册信息的信息收集单元。

公开(公告)号：KR1020140089287A

公开(公告)日：2014-07-14

申请号：KR1020130075975

申请日：2013-06-28

Applicant: 한국전자통신연구원

Inventor： 오천인 ,  이한규 ,  정영호 ,  홍진우

IPC: H04H60/46 ,  H04H60/33

CPC classification number: H04H60/33 ,  H04H60/46 ,  H04N19/146 ,  H04N19/44 ,  H04N21/251

Abstract: The present invention relates to a behavior analysis collaboration system, a method of preprocessing an image of a user terminal for behavior analysis collaboration and a method of postprocessing an image of a behavior information extraction server for behavior analysis collaboration, wherein network load can be reduced and time and cost required for user behavior analysis can be reduced by a user terminal and a server separately performing image processing for user behavior analysis without performing a one-sided user behavior analysis by one of the user terminal or the server.

Abstract translation: 本发明涉及一种行为分析协作系统，一种预处理用于行为分析协作的用户终端的图像的方法以及用于行为分析协作的行为信息提取服务器的后处理方法，其中可以减少网络负载， 用户行为分析所需的时间和成本可以由用户终端和服务器分别执行用于用户行为分析的图像处理而不执行用户终端或服务器之一的单面用户行为分析来减少。

公开(公告)号：KR1020130082723A

公开(公告)日：2013-07-22

申请号：KR1020110135375

申请日：2011-12-15

Applicant: 한국전자통신연구원

Inventor： 이상윤 ,  오천인 ,  이한규

IPC: G06Q30/02 ,  G06Q30/06

CPC classification number: G06Q30/02 ,  G06Q30/0251

Abstract: PURPOSE: A customized advertisement providing device and a method thereof are provided to recognize a passenger into a moving object and offer customized advertisements to the passenger. CONSTITUTION: An advertisement providing device (104) includes a passenger recognizing unit (201), an advertisement providing unit (202), and a page providing unit (203). The passenger recognizing unit recognizes a passenger identifier using a sensor attached to a moving object. The advertisement providing unit receives advertisements from advertisement exposure devices around the moving route of the moving object and offers the advertisements matched to the recognized passenger. The page providing unit provides a related purchase page to products in the advertisements. [Reference numerals] (104) Advertisement providing device; (201) Passenger recognizing unit; (202) Advertisement providing unit; (203) Page providing unit

Abstract translation: 目的：提供定制的广告提供装置及其方法，以将乘客识别为移动物体，并向乘客提供定制广告。 广告提供装置（104）包括乘客识别单元（201），广告提供单元（202）和页面提供单元（203）。 乘客识别单元使用附接到移动物体的传感器来识别乘客标识符。 广告提供单元从广告曝光装置接收移动物体的移动路线周围的广告，并提供与识别的乘客相匹配的广告。 页面提供单元向广告中的产品提供相关购买页面。 （附图标记）（104）广告提供装置; （201）乘客识别单位; （202）广告提供单位; （203）页面提供单元

公开(公告)号：KR1020130082721A

公开(公告)日：2013-07-22

申请号：KR1020110135364

申请日：2011-12-15

Applicant: 한국전자통신연구원

Inventor： 김재호 ,  이한규 ,  홍진우

IPC: G06Q30/02 ,  G06Q50/10 ,  G10L15/00

CPC classification number: G06Q30/0257 ,  G06Q50/10 ,  G10L15/00

Abstract: PURPOSE: Content targeting advertisement service apparatus based on user interests is provided to service advertisement content related to the interests of a smart TV user who has internet connection, while the user is watching the broadcast content. CONSTITUTION: Before broadcast content is sent, if a user directly logs in or the face of the user is recognized (S500), broadcast service apparatus sends the ID of the user to an advertisement agent (S505). Once the user selects broadcast content (S510), the broadcast service apparatus sends the ID of the broadcast content and metadata to the advertisement agent (S515). The broadcast service apparatus outputs the selected broadcast content (S520), and the user enters inputs directly through the voice recognition interface of the broadcast service apparatus or by pointing on a screen using a user terminal which is a pointing interface (S525). The broadcast service apparatus sends the information about the pointed coordinates on the user terminal, the pointing time, the voice recognition result, and the voice recognition time to the advertisement agent (S530). The advertisement agent searches for scene metadata corresponding to the voice recognition time or the pointing time from the broadcast content metadata and draws user interest analysis information using the scene metadata, the pointed coordinates, or the voice recognition result (S535). The advertisement agent transmits the user interest analysis information to the advertisement platform (S540). The advertisement platform searches for advertisement content metadata and selects the most recommendable advertisement based on the user interest analysis information (S545). The advertisement platform sends the recommendable advertisement to the broadcast service apparatus (S550). [Reference numerals] (AA) Advertisement platform; (BB) Advertisement agent; (CC) Broadcast service device; (DD) Log in or face recognition; (S505) User ID; (S510) Select broadcast contents; (S515) Broadcast contents ID, metadata; (S520) Play a broadcast content; (S525) Voice recognition or pointing reception; (S530) Voice recognition result, time or pointing coordinate, time; (S535) Search image metadata and extract user interest information; (S540) User interest information; (S545) Search advertisement contents metadata and extract recommendable advertisement; (S550) Recommendable advertisement

Abstract translation: 目的：在用户正在观看广播内容的同时，提供基于用户兴趣的内容定向广告服务设备，用于与具有因特网连接的智能电视用户的兴趣相关的服务广告内容。 规定：在发送广播内容之前，如果用户直接登录或用户的脸部被识别（S500），广播服务设备将广告代理的用户ID发送给广告代理（S505）。 一旦用户选择广播内容（S510），广播服务装置将广播内容和元数据的ID发送给广告代理（S515）。 广播服务装置输出选择的广播内容（S520），并且用户直接通过广播服务装置的语音识别接口输入输入，或者通过使用作为指示界面的用户终端指向屏幕（S525）。 广播服务装置向广告代理发送关于用户终端上的尖的坐标的信息，指示时间，语音识别结果和语音识别时间（S530）。 广告代理人从广播内容元数据中搜索对应于语音识别时间或指向时间的场景元数据，并使用场景元数据，尖点坐标或语音识别结果绘制用户兴趣分析信息（S535）。 广告代理将用户兴趣分析信息发送到广告平台（S540）。 广告平台基于用户兴趣分析信息搜索广告内容元数据并选择最可推荐的广告（S545）。 广告平台向广播服务装置发送推荐广告（S550）。 （参考编号）（AA）广告平台; （BB）广告代理; （CC）广播服务设备; （DD）登录或面部识别; （S505）用户ID; （S510）选择广播内容; （S515）广播内容ID，元数据; （S520）播放广播内容; （S525）语音识别或指点接收; （S530）语音识别结果，时间或指点坐标，时间; （S535）搜索图像元数据并提取用户兴趣信息; （S540）用户兴趣信息; （S545）搜索广告内容元数据并提取推荐广告; （S550）推荐广告

公开(公告)号：KR1020130065766A

公开(公告)日：2013-06-20

申请号：KR1020110125713

申请日：2011-11-29

Applicant: 한국전자통신연구원

Inventor： 오천인 ,  이상윤 ,  이한규 ,  홍진우

IPC: G06Q30/02

CPC classification number: G06Q30/0241

Abstract: PURPOSE: An N-screen-based PPL advertisement information providing method is provided to search for or watch broadcast content during independent broadcasting from a TV by providing PPL advertisement organization information to an N-screen terminal. CONSTITUTION: An N-screen terminal device(220) receives PPL advertisement organization list information from a broadcast content output device(210). The N-screen terminal device synchronizes the PPL advertisement organization list information with corresponding broadcast content based on advertisement time information and displays the same. The N-screen terminal device includes advertisement names about one or more products included in the PPL advertisement organization list information and storage addresses of advertisement files and displays product information about the products. [Reference numerals] (210) Broadcast content output device; (220) N-screen terminal device; (AA) PPL advertisement organization list; (BB, EE, GG) Yes; (CC, DD, FF) No; (HH) End; (S200) Transmit information about a PPL advertisement organization list according to programs; (S210) Store the PPL advertisement organization list information; (S220) Play broadcast contents; (S230) Finish the broadcast contents; (S240,S260) Watch, search, and purchase advertisements; (S250) Display product information according to time information; (S270) Store in an interest advertisement list; (S280) Manage advertisement access history of each user

Abstract translation: 目的：提供一种基于N屏幕的PPL广告信息提供方法，用于通过向N屏幕终端提供PPL广告组织信息来从TV独立广播中搜索或观看广播内容。 构成：N屏幕终端装置（220）从广播内容输出装置（210）接收PPL广告组织列表信息。 N屏幕终端设备基于广告时间信息将PPL广告组织列表信息与相应的广播内容进行同步并显示。 N屏幕终端设备包括关于包含在PPL广告组织列表信息中的一个或多个产品的广告名称和广告文件的存储地址，并且显示关于产品的产品信息。 （附图标记）（210）广播内容输出装置; （220）N屏幕终端设备; （AA）PPL广告组织名单; （BB，EE，GG）是的; （CC，DD，FF）否; （HH）结束; （S200）根据节目发送关于PPL广告组织列表的信息; （S210）存储PPL广告组织列表信息; （S220）播放广播内容; （S230）完成广播内容; （S240，S260）观看，搜索和购买广告; （S250）根据时间信息显示产品信息; （S270）存储在利息广告列表中; （S280）管理每个用户的广告访问历史

公开(公告)号：KR1020110113850A

公开(公告)日：2011-10-19

申请号：KR1020100033159

申请日：2010-04-12

Applicant: 한국전자통신연구원

Inventor： 최지훈 ,  이한규 ,  김진웅

IPC: G06Q30/02

CPC classification number: G06Q30/0241 ,  G06F17/30595

Abstract: 매쉬업 광고 제공 시스템 및 방법이 개시된다. 매쉬업 광고 제공 시스템은 광고 간에 유사성 또는 연관성 중 적어도 하나에 기초하여, 연결정보를 구성하는 정보 구성부와, 상기 연결정보에 기초하여, 제1 광고와 연관된 제2 광고 항목을 배열하는 광고 리스트를 작성하고, 상기 제1 광고를 디스플레이하는 경우, 상기 제1 광고와 함께, 상기 광고 리스트를 디스플레이하도록 제어하는 제어부를 포함한다.

公开(公告)号：KR101052480B1

公开(公告)日：2011-07-29

申请号：KR1020080121828

申请日：2008-12-03

Applicant: 한국전자통신연구원

Inventor： 최지훈 ,  김현철 ,  이한규 ,  홍진우

IPC: H04N21/236 ,  H04N21/43 ,  H04N21/472

Abstract: 본 발명은 지상파 디지털 멀티미디어 방송 시스템에서의 방송신호 송수신장치와 그 방법에 관한 것으로, 방송신호 송신장치는 DMB 콘텐츠와 연관된 소정의 어플리케이션들을 저장하는 데이터 저장부와, 상기 저장된 소정의 어플리케이션들 중에서 사용자 인터페이스를 통해 입력되는 사용자 입력 신호에 대응하는 어플리케이션의 동기화 및 제어 정보를 생성하는 동기화 및 어플리케이션 정보 생성부와, 상기 DMB 콘텐츠의 데이터를 인코딩하여 제1 전송스트림을 생성하는 제1 부호화부와, 상기 데이터 저장부로부터 상기 사용자 입력 신호에 대응하는 어플리케이션의 데이터를 입력받아 인코딩하여 인코딩된 어플리케이션 데이터를 생성하는 제2 부호화부와, 상기 제1 전송스트림과 상기 동기화 및 제어 정보를 다중화하여 다중화된 전송스트림을 생성하는 다중화부와, 상기 다중화된 전송스트림을 제공받아 스트림모드 데이터로 변환하고, 상기 인코딩된 어플리케이션 데이터를 제공받아 패킷모드 데이터로 변환하여, 상기 스트림모드 데이터와 상기 패킷모드 데이터를 앙상블 다중화하는 송출부를 포함한다.
AV(Audio Video), 어플리케이션(Application), 다중화, 부호화

Abstract translation: 本发明涉及在地面数字多媒体广播系统中发送和接收广播信号的设备和方法，并且广播信号发送设备包括：数据存储单元，用于存储与DMB内容相关联的预定应用; 到用于产生与同步和应用信息生成部的第一传输流，用于产生一个同步和控制应用程序的信息的第一编码单元，通过数据对应于用户输入的输入信号的DMB内容的数据进行编码 第二编码器，用于从存储单元接收与用户输入信号对应的应用数据并对应用数据进行编码以生成编码的应用数据;以及复用器，用于复用第一传输流和同步和控制信息， 代 在多路复用器和服务复用传输流中接收转换为流模式数据，并将所接收的提供转换成包模式数据传输单元，配置为合奏复用分组模式数据和流模式数据编码的应用数据 它包括。