-
公开(公告)号：KR1020130072063A
公开(公告)日：2013-07-01
申请号：KR1020110139610
申请日：2011-12-21
Applicant: 한국전자통신연구원
IPC: G06T7/00
CPC classification number: G06K9/00 , G06K2209/19 , G06T7/12 , G06T7/74 , G06T2207/30164
Abstract: PURPOSE: A device and method for recognizing product components are provided to increase a mounting rate of components by estimating a pose angle and kinds using feature vectors. CONSTITUTION: An image pre-processing unit of a learning unit (100) detects a component area using a component edge by extracting each component edge from inputted component images. A feature extractor of the learning unit extracts component features in the component area, and generates the feature vectors using the component features. A recognition unit (200) inputs the feature vectors in a nerve network in advance in order to recognize the type of the component. [Reference numerals] (110) Image input unit; (120) Image pre-processing unit; (130) Feature extractor; (140) Nerve network learning unit
Abstract translation: 目的：提供用于识别产品部件的装置和方法，以通过使用特征向量估计姿态角度和种类来增加部件的安装速度。 构成：学习单元（100）的图像预处理单元通过从输入的分量图像中提取每个分量边缘，使用分量边缘来检测分量区域。 学习单元的特征提取器提取组件区域中的组件特征，并使用组件特征生成特征向量。 识别单元（200）预先输入神经网络中的特征向量以识别该组件的类型。 （附图标记）（110）图像输入单元; （120）图像预处理单元; （130）特征提取器; （140）神经网络学习单元
-
公开(公告)号：KR1020120083064A
公开(公告)日：2012-07-25
申请号：KR1020110004501
申请日：2011-01-17
Applicant: 한국전자통신연구원
IPC: G06T7/00
CPC classification number: G06T7/254 , G06K9/00228 , G06T7/13 , G06T7/269
Abstract: PURPOSE: An apparatus and a method for discriminating a real face image are provided to discriminate whether the face of an image is real face or picture face. CONSTITUTION: A motion map generating unit(104) indicates a motion generating area in a pixel unit. A face detection unit(106) detects a rectangular area in which the face is located from an image. A photo edge detection unit(108) detects the girth of a photograph from the image. A motion direction detection unit(110) detects the motion direction for the background and detected face. A photo face determination unit(112) determines photograph face.
Abstract translation: 目的：提供用于区分真实脸部图像的装置和方法，以区分图像的脸部是真实面部还是图像面部。 构成：运动图生成单元（104）表示像素单位中的运动生成区域。 面部检测单元（106）从图像中检测面部位于其中的矩形区域。 照片边缘检测单元（108）从图像检测照片的周长。 运动方向检测单元（110）检测背景和检测到的面部的运动方向。 照片面确定单元（112）确定照片面。
-
公开(公告)号：KR1020120066397A
公开(公告)日：2012-06-22
申请号：KR1020100127723
申请日：2010-12-14
Applicant: 한국전자통신연구원
CPC classification number: G06K9/325 , G06K2209/15 , G06T7/00 , G06F17/30286
Abstract: PURPOSE: A method for extracting text area and an extraction apparatus are provided to accurately extract character area from an image with low resolution and noise. CONSTITUTION: A method for extracting text area comprises: a step of generating text area prediction value within a second image based on database with geometric information to the text area of a first image(110); a step of determining text recognition to a text enable area within the second image(120); and a step of selecting a text area within the second image(130).
Abstract translation: 目的：提供一种提取文本区域的方法和提取设备，以从低分辨率和噪声的图像中准确提取字符区域。 一种用于提取文本区域的方法包括：基于具有几何信息的数据库在第二图像内生成文本区域预测值到第一图像（110）的文本区域的步骤; 确定第二图像（120）内的文本使能区域的文本识别的步骤; 以及选择所述第二图像（130）内的文本区域的步骤。
-
公开(公告)号：KR1020120064572A
公开(公告)日：2012-06-19
申请号：KR1020100125856
申请日：2010-12-09
Applicant: 한국전자통신연구원
Abstract: PURPOSE: An image analysis and correction system having a high-speed image analysis function is provided to enable police officers to rapidly deal with crime by extracting information from CCTV data. CONSTITUTION: An image analysis and correction system comprises the following: an image analyzing unit(220) extracting a domain with the movement by analyzing an input image or video, and extracting human or vehicle information from the extracted domain; a database(240) for saving the analyzed information from the image analyzing unit; and an image correction unit(230) correcting the input image or video, and checking the analysis result transferred from the image analyzing unit.
Abstract translation: 目的：提供具有高速图像分析功能的图像分析和校正系统，使警务人员能够通过从CCTV数据中提取信息来迅速处理犯罪。 构成：图像分析和校正系统包括以下：图像分析单元（220）通过分析输入图像或视频来提取移动的域，以及从提取的域中提取人或车辆信息; 数据库（240），用于从所述图像分析单元保存所分析的信息; 以及校正输入图像或视频的图像校正单元（230），并且检查从图像分析单元传送的分析结果。
-
公开(公告)号：KR1020120023485A
公开(公告)日：2012-03-13
申请号：KR1020100122186
申请日：2010-12-02
Applicant: 한국전자통신연구원
CPC classification number: G06K9/6256 , G06K9/3258
Abstract: PURPOSE: A vehicle identifier recognition system and method thereof are provided to effectively extract the number of a vehicle by determining an identifier type. CONSTITUTION: An ROI(Region Of Interest) extraction unit(110) extracts ROI from an input image based on training images. An identifier region selecting unit(120) selects an identifier area including the identifier of a vehicle according to the ROI. An identifier recognition unit(130) recognizes the identifier of the vehicle by analyzing the identifier area according to the type of the identification area. A training image analysis unit(180) creates a pre-learning result based on a common data by analyzing training images.
Abstract translation: 目的：提供一种车辆识别识别系统及其方法，通过确定一种标识符类型来有效地提取车辆的数量。 构成：ROI（感兴趣区域）提取单元（110）基于训练图像从输入图像提取ROI。 标识符区域选择单元（120）根据ROI选择包括车辆的标识符的标识符区域。 标识符识别单元（130）根据识别区域的类型，通过分析标识符区域来识别车辆的标识符。 训练图像分析单元（180）通过分析训练图像来创建基于公共数据的预学习结果。
-
Patent Agency Ranking
公开(公告)号：KR1020110003820A
公开(公告)日：2011-01-13
申请号：KR1020090061275
申请日：2009-07-06
Applicant: 한국전자통신연구원
Inventor： 김도형 , 장민수 , 박천수 , 이재연 , 강상승 , 김재홍 , 김혜진 , 박찬규 , 윤영우 , 윤우한 , 윤호섭 , 이대하 , 조영조 , 조현성 , 지수영 , 손주찬 , 정연구
Abstract: PURPOSE: An electronic device driving a character profile and an operating method thereof are provided to enable a character to perform high-level interaction with a user according to a behavioral pattern of a specific character. CONSTITUTION: A capturing unit(140) captures an outside stimulus. A character platform(150) drives a character profile. A responding unit(170) expresses a reaction of a specific character about the outside stimulus. The character platform comprises a character profile slot(160) for loading the character profile. The character profile can separate from the character platform. The character profile can be replaced by another character profile.
Abstract translation: 目的：提供一种驱动字符简档及其操作方法的电子设备，以使角色能够根据特定字符的行为模式来执行与用户的高级交互。 构成：拍摄单元（140）捕获外部刺激。 字符平台（150）驱动字符简档。 响应单元（170）表示关于外部刺激的特定角色的反应。 字符平台包括用于加载字符轮廓的字符轮廓槽（160）。 角色配置文件可以与角色平台分离。 字符配置文件可以替换为另一个字符配置文件。
-
公开(公告)号：KR1020080052129A
公开(公告)日：2008-06-11
申请号：KR1020070015541
申请日：2007-02-14
Applicant: 한국전자통신연구원
Abstract: A system for acquiring speaker information by using voice feature information of a speaker and a method therefor are provided to extract the voice feature information from a voice of the speaker and acquire various information about the speaker from the extracted information. A GMM(Gaussian Mixture Model) generating unit(11) generates a GMM for inputted speaker-classified voice feature information. A speaker information deducing unit(12) deduces information about a specific speaker having a specific voice feature from the GMM and a UBM(Universal Background Model) constructed in each node of a decision tree.
Abstract translation: 提供一种通过使用讲话者的语音特征信息来获取说话者信息的系统及其方法，用于从扬声器的语音提取语音特征信息，并从提取的信息中获取关于说话者的各种信息。 GMM（高斯混合模型）生成单元（11）生成用于输入的说话者分类语音特征信息的GMM。 扬声器信息推导单元（12）从GMM中推导出具有特定语音特征的特定扬声器的信息以及在决策树的每个节点中构造的UBM（通用背景模型）。
-
公开(公告)号：KR1020080049394A
公开(公告)日：2008-06-04
申请号：KR1020060119899
申请日：2006-11-30
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00234 , G06K9/42 , G06K9/4661 , G06T7/40
Abstract: A preprocessing method for face recognition, and a method and a device for recognizing a face with the same are provided to improve a security level and performance of the face recognition by recognizing the face stably even if lighting is largely changed. An input unit(310) receives an image including a user face, and an extractor(320) extracts a face area from the inputted image. A lighting preprocessor(350) minimizes lighting effect by calculating neighboring areas of each pixel of the face area, and adjusting a pixel value according to a pixel value of a lower brightness value than the pixel in the neighboring areas. A recognizer(380) recognizes the face from a data image output from the lighting preprocessor. A geometrical preprocessor(340) normalizes the face area by keeping a ratio between positions of face components before the extracted face area is inputted to the lighting preprocessor. A database(370) stores face image data of a plurality of users having an ID. A mode controller(360) outputs the data image of the lighting preprocessor to the recognizer in a face recognition mode and stores the data image to the database in a face registration mode.
Abstract translation: 提供了一种用于人脸识别的预处理方法，以及用于识别其面部的方法和装置，以便即使照明发生很大变化也能够稳定地识别脸部的脸部识别的安全性和性能。 输入单元（310）接收包括用户面的图像，并且提取器（320）从所输入的图像中提取面部区域。 照明预处理器（350）通过计算面部区域的每个像素的相邻区域来最小化照明效果，并且根据比邻近区域中的像素更低的亮度值的像素值来调整像素值。 识别器（380）从照明预处理器输出的数据图像识别脸部。 几何预处理器（340）通过在提取的面部区域被输入到照明预处理器之前保持面部分量的位置之间的比例来对脸部区域进行归一化。 数据库（370）存储具有ID的多个用户的脸部图像数据。 模式控制器（360）以面部识别模式将照明预处理器的数据图像输出到识别器，并以面部注册模式将数据图像存储到数据库。
-
公开(公告)号：KR100826878B1
公开(公告)日：2008-05-06
申请号：KR1020060095012
申请日：2006-09-28
Applicant: 한국전자통신연구원
Abstract: 본 발명은 손 모양을 인식하는 방법 및 이를 위한 장치에 관한 것으로서, 특별한 조명 또는 배경에 대한 제한이 없고, 사용자의 연령이나, 헤어스타일, 메이크업, 사용자의 다양한 포즈 등에 무관하게, 실시간으로 손 모양을 인식할 수 있도록 하기 위하여, 사용자가 촬상된 영상으로부터 얼굴 영역을 검출하고, 상기 검출된 얼굴 영역으로부터 얼굴의 피부색 정보를 수집하여, 상기 피부색 정보에 기초하여 손 영역을 검출한 다음, 상기 검출된 손 영역에 대한 정보에 기초하여 손 모양을 분석한다.
손 모양, 인식, 인식, 피부색, 가위바위보, 묵찌빠-
公开(公告)号：KR100758707B1
公开(公告)日：2007-09-14
申请号：KR1020060039741
申请日：2006-05-03
Applicant: 한국전자통신연구원
Abstract: 본 발명은 로봇의 도움이 필요한 사람이나 개체 근처로 이동하여 주위 상황을 판단하고 적절한 대응조치를 취하기 위해 로봇으로 하여금 음원이 발생한 곳으로 이동하도록 음원 발원지의 위치 인식 및 추적을 위한 로봇 환경에서 음원 추적 시스템 및 방법에 관한 것으로서, 입력되는 음원에 대해 로봇 본체 잡음을 감소시킨 후 신호의 강도를 계산하여 강도가 큰 순서대로 두 개의 마이크로폰을 결정하고, 그 결정된 채널 사이에서 음원이 발생된 것으로 결정한 다음, 그 채널에 대해서만 시간차를 계산하여 음원의 방향 및 거리를 계산함으로써 모든 마이크로폰에 대해 시간차를 계산하는 기존의 방법보다 실행시간을 단축하며, 로봇은 음원이 발생한 위치로 이동하는 동시에 로봇의 실행시간을 고려하여 실시간으로 음원 위치를 지속적으로 확인하고 갱신하여 시스템의 신뢰도를 높이는 효과가 있다.
음원 추적, 로봇 제어, 칼만 필터, 로봇 마이크로폰, 로봇 실행시간, 시간차 추출
-
-
-
-
-
-
-
-
-
Search Results
106
records
(took 0.031 seconds)
