-
公开(公告)号：KR101472451B1
公开(公告)日：2014-12-18
申请号：KR1020100109298
申请日：2010-11-04
Applicant: 한국전자통신연구원
CPC classification number: G06F17/3069 , G06F17/30707
Abstract: 본발명은디지털콘텐츠관리시스템및 방법에대하여개시한다. 본발명의일면에따른디지털콘텐츠관리시스템은, 입력된디지털콘텐츠의특징벡터를추출하고, 상기특징벡터에컬럼부분공간맵핑을수행하여컬럼부분공간프로젝션행렬을산출하는학습모듈; 상기행렬을이용하여상기디지털콘텐츠에색인작업을수행한후, 상기행렬과상기디지털콘텐츠를저장하는색인모듈; 및상기디지털콘텐츠를검색하기위한질의데이터가입력되면, 상기질의데이터의특징벡터에컬럼부분공간맵핑을수행하고, 맵핑된상기질의데이터의특징벡터와유사도가높은상기행렬이색인하는상기디지털콘텐츠를검색하는검색모듈을포함하는것을특징으로한다.
-
公开(公告)号：KR1020140081254A
公开(公告)日：2014-07-01
申请号：KR1020120150808
申请日：2012-12-21
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00369 , G06K9/4614 , G06K9/4642 , G06T7/215 , G06T2207/10016 , G06T2207/30196
Abstract: Provided are an apparatus and a method for reliably detecting a human being from an input video image by using tilt-based feature vectors and neural networks. The provided apparatus comprises an image preprocessing unit for modeling a background image from an input image; a motion object region setting unit for differentiating the input image and the background image to set a motion object region in which a motion exists; a human being region detecting unit for extracting feature vectors of the whole body and the upper body based on a tilt from the motion object region, and inputting the extracted feature vectors of the whole body and the upper body based on a tilt to a neural network classifier to detect a human being region; and a determining unit for determining whether or not an object in the detected human being region is a human being. By using tilt-based feature vectors and neural networks having excellent discernment with respect to the whole body and the upper body of various types of images, a human being can be detected reliably.
Abstract translation: 提供了一种通过使用基于倾斜的特征向量和神经网络从输入视频图像可靠地检测人的装置和方法。 所提供的装置包括用于从输入图像对背景图像建模的图像预处理单元; 运动对象区域设定单元，用于对输入图像和背景图像进行微分，以设定存在运动的运动对象区域; 人体区域检测单元，用于基于来自运动对象区域的倾斜来提取全身和上身的特征向量，并且基于倾斜将所提取的全身和上身的特征向量输入到神经网络 分类器来检测人类区域; 以及确定单元，用于确定检测到的人区域中的对象是否是人。 通过使用相对于各种图像的全身和上身具有优异辨别性的倾斜的特征向量和神经网络，可以可靠地检测人。
-
公开(公告)号：KR101255555B1
公开(公告)日：2013-04-17
申请号：KR1020090113861
申请日：2009-11-24
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00093
Abstract: 본 발명의 지문인식 장치는, 입력되는 등록 지문 영상에 대하여 전처리과정을 특징점을 추출하는 등록 지문 특징점 추출부; 고유 정보에 기초한 다항식에 의하여 추출된 진짜 특징점을 변환하고, 추출된 진짜 특징점을 은닉하기 위한 소정의 가짜 특징점을 생성하되 진짜 특징점의 다항식 결과값과 상이한 값을 이용하며 가짜 특징점을 생성하여, 변환된 진짜 특징점과 가짜 특징점을 통합하여 등록 특징점 정보를 생성하여 저장부에 등록 저장하는 특징점 보호부; 입력되는 인증 지문 영상에서 특징점을 추출하는 인증 지문 특징점 추출부; 및 인증지문 특징점 추출부에서 추출된 특징점과 지문등록과정에서 상기 저장부에 기 저장된 등록 지문의 특징점의 유사도를 측정하여 일치 여부를 결정하는 지문 비교부를 구비한다.
지문, 인식, 특징점, 보호, 추출, 인증, 등록-
公开(公告)号：KR101173550B1
公开(公告)日：2012-08-13
申请号：KR1020090023987
申请日：2009-03-20
Applicant: 한국전자통신연구원
Abstract: 본 발명은 위조 얼굴 검출 장치에 관한 것으로, 라인 프로젝터를 이용하여 라인 빔을 일정한 간격으로 피사체에 투사하고, 적외선 카메라에 의해 촬영된 일련의 라인 빔이 투사된 피사체의 적외선 영상에 대한 위조 얼굴 특징을 추출하여 위조 얼굴을 검출함으로써, 검출결과 위조 얼굴이 아닌 경우에 대해서만 얼굴인식, 즉 사용자 인증을 수행할 수 있어 얼굴 인식 시스템의 보안성 및 신뢰성을 향상시킬 수 있다. 또한 본 발명은 위조 얼굴 검출 장치를 제공함으로써, 일반 카메라에 의해 제작된 사진과 같은 위조 얼굴과 실제 얼굴을 판별할 수 있어 얼굴 인식 시스템의 부정 사용을 사전에 방지할 수 있다.
라인 프로젝터, 적외선 카메라, 얼굴 인식, 영상, 위조-
公开(公告)号：KR1020120055404A
公开(公告)日：2012-05-31
申请号：KR1020100117134
申请日：2010-11-23
Applicant: 한국전자통신연구원
IPC: G06T7/40
CPC classification number: G06K9/00617 , G06T7/44
Abstract: PURPOSE: An apparatus and method for detecting an eye for a gaze tracking are provided to generate an average eye filter to an input face image after the average eye filter is generated by considering various eye images. CONSTITUTION: An image normalizing unit(110) normalizes an input image by using an average eye filter. An eye filtering unit filters the normalized input image by using the average eye filter. An image binarizing unit(130) binarizes the eye filtering image. An image correcting unit corrects an image error among the generated binary image. An eye detecting unit(150) detects an eye area corresponding to a correction image from an input image.
Abstract translation: 目的：提供一种用于检测眼睛注视跟踪的装置和方法，以通过考虑各种眼睛图像来生成平均眼睛滤波器之后的输入面部图像的平均眼睛过滤器。 构成：图像归一化单元（110）通过使用平均眼滤波器对输入图像进行归一化。 眼睛滤波单元通过使用平均眼滤波器对归一化的输入图像进行滤波。 图像二值化单元（130）将眼睛过滤图像二值化。 图像校正单元校正所生成的二值图像中的图像误差。 眼睛检测单元（150）从输入图像检测与校正图像相对应的眼睛区域。
-
Patent Agency Ranking
公开(公告)号：KR1020120053211A
公开(公告)日：2012-05-25
申请号：KR1020100114368
申请日：2010-11-17
Applicant: 한국전자통신연구원
CPC classification number: G06N99/005 , G06F17/30023
Abstract: PURPOSE: A multimedia data search method, an apparatus using the same, and a pattern recognition method thereof are provided to utilize learning data optimized for test data as actual learning data, thereby improving recognition/classification performance. CONSTITUTION: A database(300) stores search target data and a first search target property extracted from the search target data. A first search part(200) searches data corresponding to a first search condition property in the database by comparing the first search condition property with the first search target property after extracting the first search condition property from user input search condition data. A performance part(400) provides a machine learning process using learning data selected from a search result.
Abstract translation: 目的：提供一种多媒体数据搜索方法，使用该方法的装置及其模式识别方法，以利用针对测试数据优化的学习数据作为实际学习数据，从而提高识别/分类性能。 构成：数据库（300）存储从搜索目标数据中提取的搜索目标数据和第一搜索目标属性。 第一搜索部分（200）通过在从用户输入搜索条件数据中提取第一搜索条件属性之后比较第一搜索条件属性与第一搜索对象属性来搜索与数据库中的第一搜索条件属性相对应的数据。 演奏部分（400）使用从搜索结果中选择的学习数据提供机器学习过程。
-
公开(公告)号：KR1020120047622A
公开(公告)日：2012-05-14
申请号：KR1020100109298
申请日：2010-11-04
Applicant: 한국전자통신연구원
CPC classification number: G06F17/3069 , G06F17/30707
Abstract: PURPOSE: A digital content management system and method thereof are provided to offer stable performance in a multimedia searching process by executing a sub-space learning method regardless of numbers of registration data. CONSTITUTION: A learning module(100) extracts feature vectors of inputted digital content. The learning module calculates a column sub-space matrix by mapping sub-pace in the feature vectors. After an index module(200) makes indexes in the digital content by using the sub-space matrix, stores the digital content and the sub-space matrix. A searching module(300) searches the digital content including the indexes.
Abstract translation: 目的：提供一种数字内容管理系统及其方法，通过执行子空间学习方法来提供多媒体搜索处理中的稳定性能，而不考虑注册数据的数量。 构成：学习模块（100）提取输入的数字内容的特征向量。 学习模块通过在特征向量中映射子步骤来计算列子空间矩阵。 在索引模块（200）通过使用子空间矩阵在数字内容中进行索引之后，存储数字内容和子空间矩阵。 搜索模块（300）搜索包括索引的数字内容。
-
公开(公告)号：KR1020110051714A
公开(公告)日：2011-05-18
申请号：KR1020090108441
申请日：2009-11-11
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00281 , G06K9/00234 , G06K9/6202 , G06K9/6212
Abstract: PURPOSE: A face recognition apparatus and a method thereof are provided to offer a strong effect such as an illumination change based on a relative relation with the neighbor pixels. CONSTITUTION: An extracting unit(130) extracts a regional feature in an inputted face image. A controller(150) expresses the face image using the regional feature. A controller successively divides each region of the face image in order and generates a histogram about the divided areas. The controller recognizes a face by performing a template matching through the histogram.
Abstract translation: 目的：提供一种面部识别装置及其方法，以提供基于与相邻像素的相对关系的照明变化的强效果。 构成：提取单元（130）提取输入的面部图像中的区域特征。 控制器（150）使用区域特征来表达脸部图像。 控制器按顺序对面部图像的每个区域进行分级，并且生成关于分割区域的直方图。 控制器通过执行通过直方图的模板匹配来识别脸部。
-
公开(公告)号：KR1020110048827A
公开(公告)日：2011-05-12
申请号：KR1020090105560
申请日：2009-11-03
Applicant: 한국전자통신연구원
Abstract: PURPOSE: An authentication system is provided to realize obtaining and processing of a facial image of a user in local distance. CONSTITUTION: A camera(110) photographs an image of a user(300). An acquisition unit(120) obtains a facial image in the image. A recognition part(130) recognizes the user. If the user is recognized, a main controller(140) searches user service information. A transmission unit(150) transmits the delivered user service information to a multimedia device. A display unit(170) indicates the image.
Abstract translation: 目的：提供一种认证系统，以实现用户在本地距离的面部图像的获取和处理。 构成：相机（110）拍摄用户的图像（300）。 获取单元（120）获得图像中的面部图像。 识别部件（130）识别用户。 如果用户被识别，则主控制器（140）搜索用户服务信息。 发送单元（150）将所递送的用户服务信息发送到多媒体设备。 显示单元（170）表示图像。
-
公开(公告)号：KR1020100062809A
公开(公告)日：2010-06-10
申请号：KR1020090023987
申请日：2009-03-20
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00899 , G06K9/00228 , G06K9/2018
Abstract: PURPOSE: A method for detecting a forged face and an apparatus thereof are provided to detect a forgery face by extracting the characteristics of a forgery face based on an infrared ray image of a subject onto which a line beam is projected. CONSTITUTION: A beam projection unit(10) projects a line beam onto a subject, and an image acquisition unit(20) acquires an infrared image onto which the line beam is projected. An infrared ray image processing unit(40) extracts the characteristics of a forgery face based on the linear pattern projected onto the acquired infrared ray image. A forgery face detection unit(50) compares the range induced from the characteristics of the extracted forgery face, and detects whether or not the face is forged.
Abstract translation: 目的：提供一种用于检测锻造面的方法及其装置，用于通过基于投影线束的被摄体的红外线图像提取伪造面的特性来检测伪造面。 构成：射束投影单元（10）将线束投影到被摄体上，图像获取单元（20）获取投影线束的红外图像。 红外线图像处理单元（40）基于投影到获取的红外线图像上的线状图案来提取伪造面的特性。 伪造面部检测单元（50）比较从提取的伪造面的特性引起的范围，并且检测脸部是否被锻造。
-
-
-
-
-
-
-
-
-
Search Results
110
records
(took 0.029 seconds)
