-
公开(公告)号：KR101179497B1
公开(公告)日：2012-09-07
申请号：KR1020080131783
申请日：2008-12-22
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00228
Abstract: 본 발명은 얼굴 포즈 변화에 강인한 다수의 얼굴을 검출하기 위한 얼굴 검출 장치 및 얼굴 검출 방법에 관한 것이다. 즉, 본 발명은 얼굴을 검출하고자 하는 입력 영상에 대해 조명의 영향을 최소화하기 위해서 흑백 영상으로 변환시키고, 조명 처리 등의 전처리를 수행한 후, 어댑티브 부스팅 방법으로 off-line에서 학습시킨 얼굴과 눈의 강분류기를 이용하여, 입력 영상을 일정 비율로 축소하면서 얼굴 검출, 눈 검출, 정규화 및 얼굴 검증 단계를 실시하고, 얼굴 검증에서 통과되는 얼굴의 중심좌표, 얼굴의 너비, 얼굴의 높이, 눈의 중심좌표를 하나의 얼굴 레코드로서 생성한다. 이어 이와 같이 얻어진 얼굴 레코드에서 얼굴 중심좌표와 눈 중심좌표에 클러스터링 분류 기법을 적용하여 몇 개의 얼굴 후보군 및 눈 후보군으로 분류하고, 각각의 클러스터를 분석하여 최종 얼굴과 눈 위치를 검출한다.
얼굴, 검출, 얼굴 레코드, 클러스터링-
公开(公告)号：KR1020110133271A
公开(公告)日：2011-12-12
申请号：KR1020100052907
申请日：2010-06-04
Applicant: 한국전자통신연구원
Abstract: PURPOSE: A detection method of an iris shape is provided to calculate iris characteristic pattern using a filter which includes a second and a third standard deviation. CONSTITUTION: An iris recognition apparatus normalizes an iris image(S110). The iris recognition apparatus calculates an iris feature pattern from a normalized track(S120). The iris recognition apparatus calculates an iris characteristic vector from each iris characteristic pattern(S130). The iris recognition apparatus identifies a user based on the calculated iris feature vectors(S140).
Abstract translation: 目的：提供虹膜形状的检测方法，以使用包括第二和第三标准偏差的滤波器来计算虹膜特征图案。 构成：虹膜识别装置对虹膜图像进行归一化（S110）。 虹膜识别装置从归一化轨迹计算光圈特征图案（S120）。 虹膜识别装置从每个虹膜特征图案计算光圈特征矢量（S130）。 虹膜识别装置基于所计算的虹膜特征向量识别用户（S140）。
-
公开(公告)号：KR1020110090731A
公开(公告)日：2011-08-10
申请号：KR1020100051996
申请日：2010-06-01
Applicant: 한국전자통신연구원 , 목포대학교산학협력단
CPC classification number: G06K9/00906 , G06K9/00335 , G06T7/254
Abstract: PURPOSE: A fake video detecting apparatus and a method thereof are provided to improve the reliability and a security level. CONSTITUTION: An apparatus for determining a fake image comprises an image-acquiring block(102) for acquiring an image captured by and input from a camera, a background-learning block(104) for learning a background of the image to create a learning background, a face extracting-block(108) for extracting a face region of a person to be authenticated when an input image for authentication is transmitted from the camera, and an inter-background comparing block(110) for comparing a present background of an input image with the learning background.
Abstract translation: 目的：提供一种假影像检测装置及其方法，以提高可靠性和安全性。 一种用于确定假图像的装置，包括用于获取由相机拍摄并从相机输入的图像的图像获取块（102），用于学习图像的背景以创建学习背景的背景学习块（104） ，用于当从所述相机发送用于认证的输入图像时提取要认证的人的面部区域的面部提取块（108），以及用于比较输入的当前背景的背景间比较块（110） 形象与学习背景。
-
公开(公告)号：KR1020110057449A
公开(公告)日：2011-06-01
申请号：KR1020090113861
申请日：2009-11-24
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00093
Abstract: PURPOSE: A device and method of recognizing a finger print recognition which reinforced a security are provided to safely protect fingerprint information stored in a storing device from an external aggressor by adding fake characteristic information to real characteristic information. CONSTITUTION: A register fingerprint characteristic extracting unit(100) extracts a real characteristic from a register fingerprint image which passed through a preprocessing process. A characteristic protection unit(200) converts the real characteristic by polynomial based on identity information. The characteristic protection unit uses difference value from a polynomial result value of the real characteristic to generate fake characteristic.
Abstract translation: 目的：提供一种识别加强安全性的指纹识别的装置和方法，通过向真实特征信息添加假特征信息，将存储装置中存储的指纹信息与外部侵略者进行安全保护。 构成：寄存器指纹特征提取单元（100）从通过预处理过程的寄存器指纹图像中提取实际特性。 特征保护单元（200）基于身份信息将实数特性转换为多项式。 特征保护单元使用来自实际特征的多项式结果值的差值产生假特征。
-
公开(公告)号：KR1020110052962A
公开(公告)日：2011-05-19
申请号：KR1020090109719
申请日：2009-11-13
Applicant: 한국전자통신연구원
IPC: G06T7/40
CPC classification number: G06T7/13 , G06K9/4661 , G06T5/40 , G06T2207/20021
Abstract: PURPOSE: An apparatus and method of recognizing an object are provided to improve the reliability of an object recognition. CONSTITUTION: A control unit of a recognition apparatus receives a face image from a camera(S100). A search unit performs a search about the face image by block units(S200). The control unit extracts a feature vector based on edge histogram about the each block which is searched by the search unit(S300). A matching unit matches a pattern using the extracted feather vector(S400). The control unit recognizes a face region from the unit block(S500).
Abstract translation: 目的：提供识别对象的装置和方法，以提高对象识别的可靠性。 构成：识别装置的控制单元从摄像机接收面部图像（S100）。 搜索单元以块为单位执行关于脸部图像的搜索（S200）。 控制单元基于由搜索单元搜索的每个块的边缘直方图提取特征向量（S300）。 匹配单元使用提取的羽毛矢量匹配图案（S400）。 控制单元从单元块识别面部区域（S500）。
-
Patent Agency Ranking
公开(公告)号：KR100842266B1
公开(公告)日：2008-06-30
申请号：KR1020060118573
申请日：2006-11-28
Applicant: 한국전자통신연구원
Abstract: 본 발명에 의한 바이오 인증을 위한 보조보안장치는 사용자로부터 사용자 식별정보 및 바이오 인증 단말기를 제어하기 위한 명령어 중 어느 하나에 해당하는 문자열을 입력받는 입력부 및 바이오 인증 단말기와의 통신채널을 설정하고 바이오 인증 단말기에 명령어를 전송하는 제어부를 가진다. 본 발명에 의하면 바이오 인증 단말기와는 별도로 물리적으로 독립된 2차적인 보조보안장치를 제공함으로써 바이오 인증 단말기에 한단계 높은 보안 관리를 제공한다.
-
公开(公告)号：KR1020080048320A
公开(公告)日：2008-06-02
申请号：KR1020060118573
申请日：2006-11-28
Applicant: 한국전자통신연구원
CPC classification number: G06F21/32 , H04M1/72527
Abstract: A sub security device for biometric authentication is provided to raise a security management level of a biometric authentication terminal by providing the secondary sub security device independent of the biometric authentication terminal and connected to the biometric authentication terminal. An input unit(210) receives a character string corresponding to at least one of user ID information and a command for controlling a biometric authentication terminal from the user. A processor(220) sets a communication channel for the biometric authentication terminal and transfers the command to the biometric authentication terminal. The processor includes a communication channel setter(221) setting the communication channel through mutual authentication with the biometric authentication terminal, and a command processor(222) transferring the command to the biometric authentication terminal when an instruction included in the command is within a user authority. A display driver(230) displays the character string and/or a communication state with the biometric authentication terminal. A communication port controller(240) controls a communication port for communicating with the biometric terminal.
Abstract translation: 提供用于生物认证的子安全设备，通过提供独立于生物认证终端并连接到生物认证终端的辅助子安全设备来提高生物认证终端的安全管理级别。 输入单元（210）从用户接收与用户ID信息和用于控制生物体认证终端的命令中的至少一个相对应的字符串。 处理器（220）设置用于生物认证终端的通信信道，并将该命令传送到生物认证终端。 处理器包括通过与生物认证终端的相互认证来设置通信信道的通信信道设置器（221），以及当命令中包含的指令在用户权限内时将命令传送到生物认证终端的命令处理器（222） 。 显示驱动器（230）与生物认证终端显示字符串和/或通信状态。 通信端口控制器（240）控制用于与生物测定终端进行通信的通信端口。
-
公开(公告)号：KR1020080046490A
公开(公告)日：2008-05-27
申请号：KR1020060116016
申请日：2006-11-22
Applicant: 한국전자통신연구원
CPC classification number: G06F17/30271 , G06K9/00295 , G06K9/46 , G06T2207/30201
Abstract: A method and a device for recognizing a face by using a montage face image are provided to search the face of a criminal easily by generating the montage face image with a sketch operator based on a face image database, extracting a feature vector from a similar montage face image, and performing montage face recognition. A feature vector storing part(310) extracts feature vectors from a similar montage face image and stores the feature vectors. An authenticator(320) performs authentication by receiving a target face image, extracting feature vectors from the target face image, and comparing the extracted feature vectors with feature vectors stored in the feature vector database. The authenticator includes a second generator(321) converting the target face image into the similar montage face image, a second extractor(323) extracting feature vectors from the similar montage face image, and a similarity comparator(333) evaluating similarity by comparing feature vectors of the feature vector storing part with feature vectors of the second extractor. The second generator performs conversion by applying a sketch operation including a DIP(Difference of Inverse Probability), BDIP(Block DIP), or morphology filter.
Abstract translation: 提供一种通过使用蒙太奇面部图像来识别面部的方法和装置，用于通过基于面部图像数据库的草图算子生成蒙太奇脸部图像，轻松地搜索犯罪者的脸部，从类似的蒙太奇中提取特征向量 面部图像，并执行蒙太奇脸部识别。 特征向量存储部（310）从类似的蒙太奇面部图像中提取特征向量并存储特征向量。 认证器（320）通过接收目标人脸图像，从目标人脸图像中提取特征向量，并将所提取的特征向量与存储在特征向量数据库中的特征向量进行比较来进行认证。 认证器包括将目标人脸图像转换成类似蒙太奇脸部图像的第二发生器（321），从相似蒙太奇脸部图像提取特征向量的第二提取器（323）和通过比较特征向量来评估相似度的相似性比较器 具有第二提取器的特征向量的特征向量存储部分。 第二发生器通过应用包括DIP（反向概率的差异），BDIP（块DIP）或形态滤波器的草图操作来执行转换。
-
公开(公告)号：KR1020070105528A
公开(公告)日：2007-10-31
申请号：KR1020060037758
申请日：2006-04-26
Applicant: 한국전자통신연구원
CPC classification number: G06K9/00275 , G06K9/00288 , G06K9/522
Abstract: A method and a device for authenticating a user with a face image are provided to authenticate the user by extracting a feature vector based on a characteristic that energy is concentrated around transformation coefficients in the face image and comparing similarity of the extracted feature vector with the similarity of the registered feature vectors. A frequency transformer(110) transforms the face image of a normalized face area into frequency domain. An effective transformation coefficient extractor(120) extracts effective transformation coefficients from data in the frequency domain through energy concentration area information. A feature vector extractor(130) extracts the feature vector from the extracted effective transformation coefficients. A user authenticator(150) performs user authentication by comparing the extracted feature vector with the registered feature vector. An energy concentration area extractor extracts the energy concentration area from a plurality of training face images to extract the effective transformation coefficients.
Abstract translation: 提供一种用于认证具有面部图像的用户的方法和装置，以通过基于能量集中在面部图像中的变换系数的特性提取特征向量来认证用户，并且将提取的特征向量与相似性的相似性进行比较 的注册特征向量。 变频器（110）将归一化面部区域的面部图像变换为频域。 有效的变换系数提取器（120）从频域数据通过能量集中区域信息提取有效的变换系数。 特征向量提取器（130）从所提取的有效变换系数中提取特征向量。 用户认证器（150）通过将所提取的特征向量与所登记的特征向量进行比较来执行用户认证。 能量集中区域提取器从多个训练面部图像中提取能量集中区域，以提取有效的变换系数。
-
公开(公告)号：KR1020060104380A
公开(公告)日：2006-10-09
申请号：KR1020050026506
申请日：2005-03-30
Applicant: 한국전자통신연구원
IPC: H04N7/18
CPC classification number: H04N7/181 , G06K9/00771 , H04N5/23245 , H04N5/23296
Abstract: 시각처리 기반의 사용자 인터페이스 및 보안/감시 시스템 또는 지능형 로봇에서 두 대의 카메라를 이용한 이동 물체 추적 장치 및 그 방법이 개시되어 있다. 본 발명은 두 대의 카메라를 이용하여 글로벌 뷰우 영상에서 이동 물체를 검출하고 로칼 뷰우에 의해 상세한 영상을 획득한다. 따라서 글로벌 뷰우를 이용하여 카메라의 움직임을 최소화하고 배경 영상의 제거 등 일차적인 방법으로 목표물을 효율적으로 검출할 수 있다. 또한 로칼 뷰우와 글로벌 뷰우의 모드 상호 전환이나 글로벌 뷰우를 로칼 뷰우의 PT 좌표와 동기화 등을 통해 목표물의 이동 영역이나 위치를 쉽게 추적할 수 있다.
-
-
-
-
-
-
-
-
-
Search Results
104
records
(took 0.026 seconds)
