-
11.发明公开
公开(公告)号：KR1020110057595A
公开(公告)日：2011-06-01
申请号：KR1020090114058
申请日：2009-11-24
Applicant: 포항공과대학교 산학협력단
CPC classification number: G06K9/46 , G06K9/00281 , G06K9/4642 , G06K9/6202 , G06K2009/4666
Abstract: PURPOSE: A pattern recognizing method, apparatus thereof, and recording medium thereof which uses local LBP are provided to increase recognition performance and ratio by expressing an image through a few codes between classes. CONSTITUTION: A learning face image converts a training face image into an LBP(Local Binary Pattern) and generates a feature vector based on a class label vector(101). A frequency feature vector is calculated by showing the frequency of LBP codes of the learning face image. An OLBP(Optimal Local Binary Pattern) code for maximizing the class label vector is selected. The face image is registered by a template feature vector.
Abstract translation: 目的：提供使用本地LBP的图案识别方法，装置和记录介质，以通过在类之间的几个代码表达图像来增加识别性能和比率。 构成：学习面部图像将训练面部图像转换为LBP（局部二进制图案），并基于类别标签向量（101）生成特征向量。 通过显示学习面部图像的LBP代码的频率来计算频率特征矢量。 选择用于最大化类标签向量的OLBP（最优局部二进制模式）代码。 脸部图像由模板特征向量注册。
-
公开(公告)号：KR1020100013855A
公开(公告)日：2010-02-10
申请号：KR1020080075571
申请日：2008-08-01
Applicant: 포항공과대학교 산학협력단
Abstract: PURPOSE: A method for tracking a moving object on multiple cameras using camera hand-off is provided to increase tracking performance about a moving object using a probabilistic camera hand-off method which does not need a complex preprocessing procedure. CONSTITUTION: A block about an image of a moving object is detected(102). Proximity probability of each camera about the object is calculated(104). A camera with the maximum proximity probability is selected as a major camera(106). Hand-off is performed in a camera with higher proximity probability if major camera probabilities are different(108). A route of the object is presumed(110).
Abstract translation: 目的：提供使用照相机切换跟踪多个摄像机上的移动物体的方法，以使用不需要复杂预处理过程的概率摄像机切换方法来增加对移动物体的跟踪性能。 构成：检测关于移动物体的图像的块（102）。 计算每个摄像机关于物体的近似概率（104）。 选择具有最大接近概率的相机作为主要相机（106）。 如果主摄像机概率不同，则在具有较高接近概率的摄像机中进行切换（108）。 推测对象的路线（110）。
-
公开(公告)号：KR1020090075536A
公开(公告)日：2009-07-08
申请号：KR1020080001428
申请日：2008-01-04
Applicant: 포항공과대학교 산학협력단
CPC classification number: G06T7/277 , G06K9/00214 , G06K9/621 , G06T7/251
Abstract: A head tracing method in a particle filter using an ellipsoid model performing fast tracking a particle for the predicted place is provided to trace the operation head through the small number of particles by generating particles and estimating the movement. An ellipsoid model is initialized(100). A movement is predicted by using an adaptive state transition model(110). Particles are generated through prediction(120). The most good particle is determined. A full 3D motion recovery process is utterly performed in the most good particle. The observation model is updated.
Abstract translation: 提供了使用执行快速跟踪预测位置的粒子的椭圆体模型的粒子滤波器中的头跟踪方法，以通过产生粒子并估计运动来跟踪操作头通过少量的粒子。 椭圆体模型被初始化（100）。 通过使用自适应状态转换模型（110）来预测运动。 通过预测生成粒子（120）。 确定最好的颗粒。 完全3D运动恢复过程完全在最好的粒子中执行。 观察模型更新。
-
公开(公告)号：KR1020160033552A
公开(公告)日：2016-03-28
申请号：KR1020140124637
申请日：2014-09-18
Applicant: 한화테크윈 주식회사 , 포항공과대학교 산학협력단
CPC classification number: G06K9/00288 , G06K9/4671 , G06K9/6215
Abstract: 본발명에따른키포인트기술자매칭및 다수결기법기반얼굴인식시스템및 방법에의하면, 입력된얼굴영상과등록얼굴영상으로부터키포인트기술자를추출하여비교하고, 다수결기법으로동일인여부를판단함으로써, 더빠르고정확한얼굴인식이가능하다.
Abstract translation: 根据本发明的基于关键点描述符匹配和多数表决方法的面部识别系统和面部识别方法，从输入面部图像和登记的面部图像中提取关键点描述符，并将关键点描述符和 使用多数投票方法确定被识别的脸部是否是注册用户的脸部，从而可以更快地更准确地执行脸部识别。 面部识别系统包括组合面部生成单元，关键点描述符提取单元和用于确定识别的面部是否是注册用户的脸部的登记面部确定单元。
-
公开(公告)号：KR101360412B1
公开(公告)日：2014-02-27
申请号：KR1020080090002
申请日：2008-09-11
Applicant: 현대자동차주식회사 , 포항공과대학교 산학협력단
Abstract: 본 문서는 적외선 카메라를 이용하여 영상을 얻고 운전자의 고개 움직임을 분석하여 운전자의 산만 운전이나 졸음 운전을 판별하고 이를 운전자에게 알림으로써 안전 운전에 도움을 주는 시스템이다. 특히 선글라스나 안경을 쓴 운전자의 경우와 같이 눈이 검출되지 않는 경우에는 눈이 아닌 입과 코의 위치 정보를 이용하여 운전자의 머리 움직임 정보를 파악하기 때문에, 기존의 산만/졸음 운전 감시 시스템들보다 폭넓은 운전자층을 대상으로 적용할 수 있는 자동차 안전 운전 시스템이다.
머리 움직임, 산만/졸음 운전 판단, AAM-
Patent Agency Ranking
公开(公告)号：KR1020140010541A
公开(公告)日：2014-01-27
申请号：KR1020120076452
申请日：2012-07-13
Applicant: 삼성전자주식회사 , 포항공과대학교 산학협력단
CPC classification number: G06K9/46 , G06F1/1686 , G06F3/013 , G06K9/0061 , G06K9/4652 , G06K9/605 , G06T7/11 , G06T7/143 , G06T7/149 , G06T11/60 , G06T15/04 , G06T19/00 , G06T2207/20124 , G06T2207/30041 , G06T2210/44 , H04N5/2351
Abstract: In the present invention, a method for correcting a user's staring direction in an image comprises the steps of: setting eye outline points which define areas of the user's eyes in an original image; converting the set eye outline points to a staring direction of a reference camera set previously; and converting the eye areas of the original image according to the converted outline points. [Reference numerals] (S310) Inputting images; (S320) Detecting face and eyes; (S330) Detecting eye outline points; (S340) Converting the eye outline points to a staring direction of a camera; (S350) Warping texture
Abstract translation: 在本发明中，一种用于校正用户在图像中的凝视方向的方法包括以下步骤：设置定义用户眼睛在原始图像中的区域的眼睛轮廓点; 将设定的眼睛轮廓点转换成先前设定的参考相机的起始方向; 并根据转换的轮廓点转换原始图像的眼睛区域。 （附图标记）（S310）输入图像; （S320）检测脸部和眼睛; （S330）检测眼睛轮廓点; （S340）将眼睛轮廓点转换为照相机的起始方向; （S350）翘曲纹理
-
公开(公告)号：KR1020120007959A
公开(公告)日：2012-01-25
申请号：KR1020110059649
申请日：2011-06-20
Applicant: 한화테크윈 주식회사 , 포항공과대학교 산학협력단
CPC classification number: G06T7/11 , G06T7/44 , G06T2207/10152
Abstract: PURPOSE: A method for detecting a shape by a camera is provided to be robust against a light change by using a texture pattern without directly using a gradation pattern. CONSTITUTION: The gradation difference values between the gradation of a central pixel and the gradation of surrounding pixels are calculated for each local area of an image frame(S32). The average values of the gradation difference values are compared with the gradation difference values for each local area(S33). The value of an LGP(Local Gradient Pattern) is obtained based on the comparison result. The area of a specific shape is detected from the image frame using values of the value of the LGP(S34).
Abstract translation: 目的：提供一种用于通过照相机检测形状的方法，以便通过使用纹理图案来抵抗光线变化而不直接使用渐变图案。 构成：针对图像帧的每个局部区域计算中心像素的灰度与周围像素的灰度之间的灰度差值（S32）。 将灰度差值的平均值与每个局部区域的灰度差值进行比较（S33）。 基于比较结果获得LGP（本地梯度图案）的值。 使用LGP的值的值从图像帧检测特定形状的区域（S34）。
-
公开(公告)号：KR1020100081874A
公开(公告)日：2010-07-15
申请号：KR1020090001296
申请日：2009-01-07
Applicant: 포항공과대학교 산학협력단
CPC classification number: G06K9/00302 , G06K9/6277 , G06T7/75 , G06T2207/20121
Abstract: PURPOSE: A method and a device for user-customized facial expression recognition are provided to extract strong feature point from an external factor and a noise, thereby performing independent facial expression of a person in a real time. CONSTITUTION: An image receiver receives a test image sequence from a user(S130). An image processor uses difference value between AAM(Active Appearance Model) parameters of an absence of facial expression and the test image sequence to calculate D-AAM(Differential-AAM) feature point(S140). The image processor reduces dimension by projecting the D-AAM feature point to educated manifold space(S150). The image processor recognizes facial expression of the test image sequence from the D-AAM feature points projected to the manifold space by referring to gallery sequence(S160).
Abstract translation: 目的：提供用户定制的面部表情识别的方法和装置，以从外部因素和噪声中提取强特征点，从而实时地执行人的独立的面部表情。 构成：图像接收器从用户接收测试图像序列（S130）。 图像处理器使用不存在面部表情的AAM（活动外观模型）参数与测试图像序列之间的差值来计算D-AAM（Differential-AAM）特征点（S140）。 图像处理器通过将D-AAM特征点投影到受过教育的歧管空间来减小尺寸（S150）。 图像处理器通过参照画廊序列识别来自投影到歧管空间的D-AAM特征点的测试图像序列的面部表情（S160）。
-
公开(公告)号：KR1020100030991A
公开(公告)日：2010-03-19
申请号：KR1020080090002
申请日：2008-09-11
Applicant: 현대자동차주식회사 , 포항공과대학교 산학협력단
Abstract: PURPOSE: A distracted/drowsy driving monitor system using head motion information is provided to motor drivers' distracted/drowsy driving which causes car accidents through the motion of a driver's motion. CONSTITUTION: A distracted/drowsy driving monitor system using head motion information comprises a face detection part(20), an eye detection discriminating unit(30), a head angle and position measurement part(40), a head motion tracker(50), and a distracted/drowsy driving discriminating unit(60). The face detection part extracts the face range into face image from an input image. The eye detection discriminating unit discriminates eye from the face image. The head angle and position measurement part extracts the head angle the location through the first method. The head angle and position measurement part extracts the head angle and position measurement part through a second method. The head motion tracker tracks the motion of the head using a cylinder and an ellipse model which are installed based on the heat angle and position information.
Abstract translation: 目的：使用头部运动信息的分心/昏昏欲睡的驾驶监视器系统被提供给驾驶员的分心/困倦驾驶，其通过驾驶员动作的运动引起车祸。 构成：使用头部运动信息的分心/困倦的驾驶监视器系统包括面部检测部分（20），眼睛检测判别单元（30），头部角度和位置测量部分（40），头部运动跟踪器（50） 和分心/困倦驾驶鉴别单元（60）。 面部检测部从输入图像中提取脸部范围为脸部图像。 眼睛检测鉴别单元将眼睛与脸部图像区分开。 头角和位置测量部分通过第一种方法提取头角度的位置。 头部角度和位置测量部件通过第二种方法提取头部角度和位置测量部件。 头部运动跟踪器使用基于热角度和位置信息安装的圆柱体和椭圆模型跟踪头部的运动。
-
公开(公告)号：KR1020100011280A
公开(公告)日：2010-02-03
申请号：KR1020080072424
申请日：2008-07-24
Applicant: 포항공과대학교 산학협력단
CPC classification number: G06K9/00302 , G06T3/0006 , G06T7/246
Abstract: PURPOSE: A method and an apparatus for recognizing detailed facial expression using facial expression information amplification are provided to recognize detailed facial expression by changing facial expression acknowledged as a gesture amplification module into amplified facial expression. CONSTITUTION: A face feature extracting module(101) extracts a face characteristic point from continuous face image in a predetermined time interval. A motion measuring module(102) produces a motion vector of face characteristic point based on a location change of the extracted face characteristic point. A facial expression amplifying module(103) creates the face image of amplified gesture by amplified face characteristic points using an amplifying vector. A facial expression recognizing module(104) classifies the amplified face characteristic points and recognizes facial expression.
Abstract translation: 目的：提供一种用于使用面部表情信息放大识别详细面部表情的方法和装置，用于通过将确认为手势放大模块的面部表情改变为放大的面部表情来识别详细的面部表情。 构成：面部特征提取模块（101）以预定的时间间隔从持续面部图像提取面部特征点。 运动测量模块（102）基于提取的面部特征点的位置变化产生面部特征点的运动矢量。 面部表情放大模块（103）使用放大矢量通过放大的脸部特征点创建放大手势的脸部图像。 面部表情识别模块（104）对放大的面部特征点进行分类并识别面部表情。
-
-
-
-
-
-
-
-
-
Search Results
71
records
(took 0.028 seconds)
