-
公开(公告)号：KR101412727B1
公开(公告)日：2014-07-01
申请号：KR1020130139394
申请日：2013-11-15
Applicant: 동국대학교 산학협력단
CPC classification number: G06K9/00288 , G06K9/00234 , G06K9/4647 , G06K9/66
Abstract: A device for recognizing a face comprises a communications interface which receives one or more learning images and recognizing images; a processor which identifies the face of a user of a face image according to a command; and a memory which stores the command. The command performs a step for receiving a learning image and extracting a histogram feature vector for each block; a step for generating a matching value by comparing the histogram feature vector for each block for the learning image; a step for calculating a learning weight value for each block according to the matching value; a step for adjusting a parameter of a multilayered neural network according to the weight value; a step for inputting the pixel average and the standard deviation of the recognized image and a stored registration image to the multilayered neural network and calculating an applied weight value for each block; a step for calculating the final matching value between the recognized image and the registration image by applying the applied weight value; and a step for identifying the user of the recognized image as the user of the registration image when the final matching value is equal to or greater than a predetermined threshold value.
Abstract translation: 用于识别面部的装置包括：接收一个或多个学习图像和识别图像的通信接口; 根据命令识别面部图像的用户的脸部的处理器; 以及存储命令的存储器。 该命令执行用于接收学习图像并提取每个块的直方图特征向量的步骤; 通过比较用于学习图像的每个块的直方图特征向量来生成匹配值的步骤; 根据匹配值计算每个块的学习权重值的步骤; 根据权重值调整多层神经网络的参数的步骤; 将识别图像的像素平均值和标准偏差以及存储的登记图像输入到多层神经网络并计算每个块的应用权重值的步骤; 通过应用所述加权值来计算所识别的图像和所述登记图像之间的最终匹配值的步骤; 以及当最终匹配值等于或大于预定阈值时，将所识别的图像的用户识别为注册图像的用户的步骤。
-
公开(公告)号：KR101395819B1
公开(公告)日：2014-05-16
申请号：KR1020130139398
申请日：2013-11-15
Applicant: 동국대학교 산학협력단
Abstract: An eye tracking device includes: a front camera which is positioned to face the front of a user to create a front image; an eye camera which is positioned to face the eyes of the user to create an eye image; a processor which tracks the gaze of the user with reference to the eye image according to pre-defined instructions, and points an object on the front image corresponding to the gaze of the user; and a memory which stores the instructions. The instructions are defined to perform the steps: extracting first pupil coordinates from the eye image when a user gazes at the vertexes of a pre-specified rectangle; calculating the coordinates of the front image corresponding to the first pupil coordinates according to a pre-specified transformation formula; calculating an error between the coordinates of the front image and reference coordinates that are the coordinates of the vertexes; creating a corrected transformation formula by correcting the transformation formula so as to reduce the error; extracting second pupil coordinates from the eye image; calculating the coordinates of the front image corresponding to the second pupil coordinates according to the corrected transformation formula; and pointing an object corresponding to the coordinates of the front image.
Abstract translation: 眼睛跟踪装置包括：正面照相机，其被定位成面对使用者的前方以创建正面图像; 定位成面对用户的眼睛以创建眼睛图像的眼睛相机; 根据预定义的指令，参照眼睛图像跟踪用户的凝视的处理器，并且对应于用户的视线的前方图像上的对象指向; 以及存储指令的存储器。 这些指令被定义为执行以下步骤：当用户注视到预定矩形的顶点时，从眼睛图像中提取第一瞳孔坐标; 根据预定的变换公式计算与第一瞳孔坐标对应的前图像的坐标; 计算前图像的坐标和作为顶点坐标的参考坐标之间的误差; 通过校正变换公式来产生校正的变换公式以减少误差; 从眼图中提取第二瞳孔坐标; 根据校正的变换公式计算与第二瞳孔坐标对应的前图像的坐标; 并指向与前方图像的坐标对应的对象。
-
公开(公告)号：KR1020130121303A
公开(公告)日：2013-11-06
申请号：KR1020120044380
申请日：2012-04-27
Applicant: 한국전자통신연구원 , 동국대학교 산학협력단
IPC: G06T7/20
CPC classification number: G06F3/013 , G06F3/0304
Abstract: The present invention relates to a method and a system for tracking a gaze at a distance. The system for tracking a gaze at a distance according to one embodiment of the present invention comprises: an infrared lighting unit which includes a plurality of infrared lights emitting infrared light toward a user; a gaze tracking module which tracks the position of the user’s face and collects, from the tracked face position, an eye image including at least one reflected light among a plurality of corneal reflected lights caused by the emitted infrared light and a lens reflected light caused by spectacle lenses; a processor which compares the intensity of the lens reflected light in the collected eye image with a threshold and calculates a gaze position by detecting central coordinates of the corneal reflected lights if the intensity of the lens reflected light is less than or equal to the threshold. [Reference numerals] (200) System for tracking a gaze at a distance;(210) Infrared lighting unit;(220) Gaze tracking module;(230) Processor
Abstract translation: 本发明涉及一种用于跟踪远距离凝视的方法和系统。 根据本发明的一个实施例的用于跟踪距离的目视的系统包括：红外线照明单元，其包括朝向用户发射红外光的多个红外光; 注视跟踪模块，其跟踪用户脸部的位置并且从跟踪的脸部位置收集包括由所发射的红外光引起的多个角膜反射光中的至少一个反射光的眼睛图像和透镜反射光 由眼镜镜片引起; 将收集的眼图像中的透镜反射光的强度与阈值进行比较的处理器，如果透镜反射光的强度小于或等于阈值，则通过检测角膜反射光的中心坐标来计算注视位置。 （200）用于跟踪远距离凝视的系统;（210）红外线照明单元;（220）凝视跟踪模块;（230）处理器
-
公开(公告)号：KR101302661B1
公开(公告)日：2013-09-03
申请号：KR1020120066166
申请日：2012-06-20
Applicant: 동국대학교 산학협력단 , 한국기초과학지원연구원
Abstract: 신원 인식 장치는 사용자 별 등록 홍채 영상 및 등록 얼굴 영상을 저장하는 저장부; 사용자의 얼굴 촬영에 따른 얼굴 영상과 상기 등록 얼굴 영상 간의 유클리디안 거리(Euclidian distance)를 산출하는 얼굴 인식부; 상기 사용자의 홍채 촬영에 따른 홍채 영상과 상기 등록 홍채 영상 간의 해밍 거리(Hamming distance) 중 하나 이상을 산출하는 홍채 인식부; 및 상기 유클리디안 거리 및 상기 해밍 거리 중 하나 이상에 따라 상기 사용자의 신원을 인식하는 신원 인식부를 포함한다.
-
公开(公告)号：KR101300166B1
公开(公告)日：2013-08-26
申请号：KR1020110146795
申请日：2011-12-30
Applicant: 동국대학교 산학협력단 , 한국기초과학지원연구원
Abstract: 홍채 검출 장치는 제1 카메라를 통해 사용자의 홍채 영상을 생성하는 홍채 영상 생성부, 제2 카메라를 통해 상기 사용자의 얼굴 영상을 생성하는 얼굴 영상 생성부, 상기 얼굴 영상에서 눈 영역을 검출하고, 상기 눈 영역 상의 홍채 크기를 측정하고, 상기 홍채 크기에 따라 촬영 거리를 산출하는 촬영 거리 산출부, 상기 촬영 거리 및 상기 촬영 거리에 상응하는 하나 이상의 오차 촬영 거리에 따라, 상기 홍채 영상의 홍채 영역을 검출하기 위한 탐색 대상이 되는 영역인 홍채 탐색 영역을 검출하는 탐색 영역 검출부; 및 상기 홍채 탐색 영역에서 홍채를 검출하여 홍채 정보를 생성하는 홍채 검출부를 포함한다.
-
Patent Agency Ranking
公开(公告)号：KR101201915B1
公开(公告)日：2012-11-15
申请号：KR1020110021307
申请日：2011-03-10
Applicant: 동국대학교 산학협력단
CPC classification number: A61B5/0059 , A61B5/489 , A61B2560/0406 , A61B2562/046
Abstract: 본 발명은 혈관을 촬영하고자 하는 신체가 놓이는 장소를 제공하는 본체; 상기 본체의 일측에 구비되어 근적외선을 조사하는 조명부; 조명부에서 조사된 근적외선이 신체를 투과하거나 신체에 반사되며 제공되는 광을 이용하여 신체의 혈관을 촬영하는 촬영부; 및 상기 본체의 일측에 연결설치되어 상기 조명부 및 촬영부의 작동을 위한 전류를 제공하는 전원공급부를 포함하는 의료용 혈관영상 촬영장치에 관한 것이다.
-
公开(公告)号：KR1020120103189A
公开(公告)日：2012-09-19
申请号：KR1020110021307
申请日：2011-03-10
Applicant: 동국대학교 산학협력단
CPC classification number: A61B5/0059 , A61B5/489 , A61B2560/0406 , A61B2562/046
Abstract: PURPOSE: A blood vessel imaging camera for medical treatment is provided to prevent side effect due to a contrast medium by obtaining a blood vessel image without the contrast medium. CONSTITUTION: Lighting parts(4,4') are formed at one side of a main body. The lighting parts discharge near-infrared rays to a human body. An image pickup part(6) is formed on the other side of the main body. The image pickup part takes a photograph of a blood vessel of the main body. A power supply part(10) is installed at one side of the main body. The power supply unit provides current for operation of the lighting parts and the image pickup part.
Abstract translation: 目的：提供用于医疗的血管成像相机，以通过获得没有造影剂的血管图像来防止造影剂引起的副作用。 构成：照明部件（4,4'）形成在主体的一侧。 照明部件将近红外线放射至人体。 图像拾取部分（6）形成在主体的另一侧上。 图像拾取部拍摄主体的血管的照片。 电源部件（10）安装在主体的一侧。 电源单元为照明部件和图像拾取部件的操作提供电流。
-
公开(公告)号：KR101141312B1
公开(公告)日：2012-05-04
申请号：KR1020110045782
申请日：2011-05-16
Applicant: 동국대학교 산학협력단
CPC classification number: G06T5/50 , G06T5/003 , G06T2207/10048 , G06T2207/20221 , G06T2207/30101
Abstract: PURPOSE: A medical image processing method for a blood vessel based on an image fusion method is provided to obtain a plurality of clear blood vessel images from one blood vessel image using various image fusion methods. CONSTITUTION: An original blood vessel image is filtered by a Garbor filter. An edge about the original blood vessel image and a brightness change between a blood vessel and a non blood vessel portion of the filtered blood vessel image is detected. The original blood vessel image and the edge-detected blood vessel image are fused each other. After the filtered blood vessel image is fused with the edge-detected blood vessel image, a weight value is determined using a neural network. A blood vessel image becomes clear by fusing the two images.
Abstract translation: 目的：提供一种基于图像融合方法的用于血管的医学图像处理方法，以使用各种图像融合方法从一个血管图像获得多个清晰的血管图像。 构成：原始血管图像由加法尔过滤器过滤。 检测关于原始血管图像的边缘以及血管和血管图像的非血管部分之间的亮度变化。 原始血管图像和边缘检测血管图像彼此融合。 滤波后的血管图像与边缘检测血管图像融合后，使用神经网络确定体重值。 通过融合两个图像，血管图像变得清晰。
-
公开(公告)号：KR1020110076723A
公开(公告)日：2011-07-06
申请号：KR1020100026453
申请日：2010-03-24
Applicant: 한국전자통신연구원 , 동국대학교 산학협력단
IPC: G06T7/40 , G06F3/0346
CPC classification number: G06K9/00597 , G06F3/013 , G06T7/215
Abstract: PURPOSE: An eyeline tracing apparatus using a pupil detection device and a method thereof are provided to trace the eyeline of a user by detecting pupil area. CONSTITUTION: A face detecting unit(100) detects a face region among face images. A candidate detecting unit(102) detects an eye candidate region. A pupil detecting unit(104) detects eye region. The pupil detecting unit detects the pupil by binarizing the eye area. The pupil detecting unit traces the component of the detected pupil. The pupil detecting unit detects the central coordinates of the pupil.
Abstract translation: 目的：提供使用瞳孔检测装置及其方法的眼线追踪装置，通过检测瞳孔区域来追踪使用者的眼线。 构成：面部检测单元（100）检测脸部图像中的面部区域。 候选检测单元（102）检测眼睛候选区域。 瞳孔检测单元（104）检测眼睛区域。 瞳孔检测单元通过二值化眼睛区域来检测瞳孔。 瞳孔检测单元跟踪检测到的瞳孔的分量。 瞳孔检测单元检测瞳孔的中心坐标。
-
公开(公告)号：KR100940902B1
公开(公告)日：2010-02-08
申请号：KR1020090042006
申请日：2009-05-14
Applicant: 동국대학교 산학협력단
IPC: G06K9/00
CPC classification number: G06F21/32 , G06K9/001 , G06K9/52 , G06K9/6204 , G06K2009/00932 , G06T7/13
Abstract: PURPOSE: A biometrics using finger geometry information is provided to realize the finger geometry recognition equipment of the small size by the finger geometric information. CONSTITUTION: An individual finger image is obtained through camera(S100). Using the division of the edge area and finger domain, only the finger domain is extracted from the individual finger image(S200). The minute geometrical property pattern of the finger is extracted from the finger domain(S300). Based on the geometrical property pattern, the similarity of the registered private feature pattern is measured(S400). The approval or the reject of the biometric result is decided through the pattern classification algorithm(S500).
Abstract translation: 目的：提供使用手指几何信息的生物识别技术，通过手指几何信息实现小尺寸的手指几何识别设备。 构成：通过照相机获得单独的手指图像（S100）。 使用边缘区域和手指区域的划分，仅从单个手指图像中提取手指区域（S200）。 从手指区域提取手指的微小几何属性图案（S300）。 基于几何属性模式，测量了登记的私有特征模式的相似度（S400）。 通过模式分类算法（S500）确定生物特征结果的批准或拒绝。
-
-
-
-
-
-
-
-
-
Search Results
128
records
(took 0.031 seconds)
