-
公开(公告)号：KR1020130085315A
公开(公告)日：2013-07-29
申请号：KR1020120006391
申请日：2012-01-19
Applicant: 한국전자통신연구원
IPC: G08B21/02
CPC classification number: H04N7/002 , G06K9/00369 , G06K9/00771 , G08B13/19613 , G08B13/19652 , H04N7/183
Abstract: PURPOSE: A risk recognition method based on human identification is provided to offer effective security services by reducing alarm error rates. CONSTITUTION: An identification unit generates identification information in a person by identifying detected information when the detected information is received from the detection unit detecting a human from image information captured by a photographing unit (S10-S30). The identification unit extracts access control information in a person by using identification information (S40). A dangerous situation sensing unit determines dangerous situations by analyzing the access control information by receiving a moving path tracked by a tacking unit (S50-S60). The dangerous situation sensing unit selectively warns the human against the dangerous situation when the dangerous situation is generated (S70). [Reference numerals] (AA) Start; (BB) End; (S10) Capture an image; (S20) Detect a person; (S30) Identify the person; (S40) Extract access control information; (S50) Track a moving path; (S60) Dangerous situation?; (S70) Generate a danger warning; (S80) Transmit the identification information and the image information to a management system
Abstract translation: 目的：提供基于人体识别的风险识别方法，通过降低报警错误率提供有效的安全服务。 构成：识别单元通过从从拍摄单元拍摄的图像信息中检测到人的检测单元接收到检测信息时识别检测到的信息来生成识别信息（S10-S30）。 识别单元通过使用识别信息提取人的访问控制信息（S40）。 危险情况感测单元通过接收由定位单元跟踪的移动路径来分析访问控制信息来确定危险情况（S50-S60）。 当危险情况产生时，危险情况检测单元选择性地警告人类处于危险状态（S70）。 （附图标记）（AA）开始; （BB）结束; （S10）拍摄图像; （S20）检测人; （S30）识别人; （S40）提取访问控制信息; （S50）跟踪移动路径; （S60）危险情况？ （S70）产生危险警告; （S80）将识别信息和图像信息发送到管理系统
-
公开(公告)号：KR1020130067758A
公开(公告)日：2013-06-25
申请号：KR1020110134652
申请日：2011-12-14
Applicant: 한국전자통신연구원
IPC: G06T7/00
CPC classification number: G06K9/00362 , G06F17/30247 , G06K9/6269 , G06N99/005
Abstract: PURPOSE: Human detection apparatus using SVM(Support Vector Machine) learning and a method thereof are provided to generate a histogram for a binary characteristic pattern, obtain a characteristic vector, and then detect a person from an input image, thereby obtaining a stable pattern through a cumulative sum-based analysis in spite of noise in the vicinity when obtaining a characteristic pattern for each pixel. CONSTITUTION: A learning image characteristic extraction block(112) extracts a binary learning characteristic pattern from an input learning image(102) and generates a learning characteristic vector. An SVM learning block(114) generates a classifier through the SVM learning of the generated learning characteristic vector. An image characteristic extraction block(132) extracts a binary input characteristic pattern from an input image(104) and generates an input characteristic vector. A human detection judgment block(136) detects a person by using a pre-set search region and the classifier stored in a database(120). [Reference numerals] (112) Learning image characteristic extraction block; (114) SVM learning block; (116) SVM determination weighted value generation block; (132) Input image characteristic extraction block; (134) Search area setting block; (136) Human detection judgment block
Abstract translation: 目的：提供使用SVM（支持向量机）学习的人体检测装置及其方法，以生成二进制特征图案的直方图，获得特征向量，然后从输入图像中检测人，从而获得稳定的图案 当获得每个像素的特征图案时，尽管附近有噪声的累积和分析。 构成：学习图像特征提取块（112）从输入学习图像（102）提取二进制学习特征模式，并生成学习特征向量。 SVM学习块（114）通过生成的学习特征向量的SVM学习生成分类器。 图像特征提取块（132）从输入图像（104）提取二进制输入特征图案并生成输入特征向量。 人检测判断块（136）通过使用存储在数据库（120）中的预设搜索区域和分类器来检测人。 （参考数字）（112）学习图像特征提取块; （114）SVM学习块; （116）SVM确定加权值生成块; （132）输入图像特征提取块; （134）搜索区域设置块; （136）人体检测判断块
-
公开(公告)号：KR1020130047223A
公开(公告)日：2013-05-08
申请号：KR1020110112116
申请日：2011-10-31
Applicant: 한국전자통신연구원
CPC classification number: H04N5/275 , G06K9/00771 , H04N5/272
Abstract: PURPOSE: A monitoring image-based privacy area protecting method and a device thereof are provided to easily hide an important area related to privacy in an image inputted through a camera without changing an original image. CONSTITUTION: An encoding unit encodes an original image(S110). The encoded original image is transmitted to a monitoring device or stored in a storage device(S120). A decoding unit decodes the encoded image(S130). A target detecting unit detects a target object from the encoded image to detect a privacy protection area from the target object(S140). A privacy area protecting unit hides only the privacy protection area by graphic processing(S150). [Reference numerals] (AA) Start; (BB) End; (S110) Encode a collected image; (S120) Store the encoded image; (S130) Decode the encoded image; (S140) Detect/track a target object from the decoded image; (S150) Graphic process a privacy protection area of the detected target object
Abstract translation: 目的：提供一种基于监控图像的隐私区域保护方法及其装置，以便容易地隐藏通过相机输入的图像中与隐私有关的重要区域，而不改变原始图像。 构成：编码单元编码原始图像（S110）。 编码的原始图像被发送到监视设备或存储在存储设备中（S120）。 解码单元解码编码图像（S130）。 目标检测单元从编码图像检测目标物体，以从目标对象检测隐私保护区域（S140）。 隐私区域保护单元仅通过图形处理隐藏隐私保护区域（S150）。 （附图标记）（AA）开始; （BB）结束; （S110）编码收集的图像; （S120）存储编码图像; （S130）对编码图像进行解码; （S140）从解码图像检测/追踪目标对象; （S150）图形处理检测到的目标对象的隐私保护区域
-
公开(公告)号：KR1020120071226A
公开(公告)日：2012-07-02
申请号：KR1020100132885
申请日：2010-12-22
Applicant: 한국전자통신연구원
IPC: G06T7/00
CPC classification number: G06T7/254 , G06K9/34 , G06T2207/20021
Abstract: PURPOSE: An object extracting apparatus and an apparatus thereof are provided to accurately extract an object in a background image by using multi-view fusing analysis. CONSTITUTION: An image input unit(100) generates a background image from a plurality of images. A foreground candidate decision unit(200) obtains a foreground candidate pixel by using similarity of the input image and the background image. A pixel extracting unit(300) applies a preset object extracting algorithm to the foreground candidate pixel. The pixel extracting unit extracts the object from the background.
Abstract translation: 目的：提供一种物体提取装置及其装置，通过使用多视点融合分析来准确地提取背景图像中的对象。 构成：图像输入单元（100）从多个图像生成背景图像。 前景候选决策单元（200）通过使用输入图像和背景图像的相似度来获得前景候选像素。 像素提取单元（300）将预设对象提取算法应用于前景候选像素。 像素提取单元从背景中提取对象。
-
公开(公告)号：KR1020100073076A
公开(公告)日：2010-07-01
申请号：KR1020080131661
申请日：2008-12-22
Applicant: 한국전자통신연구원
IPC: G06Q50/20
Abstract: PURPOSE: An education apparatus and method using virtual reality are provided to appear lecturer of class and remote place to same virtual world and provide contents by recognition of student's gesture thereby improving feeling of immersion of the student. CONSTITUTION: A video computers(120,122) extracts video of a student from video photographed the student. A remote teacher computer(190) extracts video of the remote place lecturer from video photographed the remote place lecturer. A class teacher computer(140) performs learning contents realizing virtual space. The class teacher computer performs real-time telecommunication with the remote place teacher computer. A virtual reality computer(150) synthesizes the student video and the remote lecturer video in the virtual space.
Abstract translation: 目的：提供使用虚拟现实的教学设备和方法，将课堂和偏远地区的讲师列为同一虚拟世界，并通过识别学生的姿态提供内容，从而提高学生的沉浸感。 规定：视频电脑（120,122）从拍摄学生的视频中提取学生的视频。 远程教师计算机（190）从远程地点讲师的视频拍摄中提取远程教室讲师的视频。 课堂老师计算机（140）执行实现虚拟空间的学习内容。 课堂老师电脑与远程教师电脑进行实时通信。 虚拟现实计算机（150）在虚拟空间中合成学生视频和远程讲师视频。
-
Patent Agency Ranking
公开(公告)号：KR1020070059607A
公开(公告)日：2007-06-12
申请号：KR1020050118628
申请日：2005-12-07
Applicant: 한국전자통신연구원
CPC classification number: G06K9/0061
Abstract: A method and a device for recognizing an iris by using a CPA based on a cumulative sum are provided to authenticate a user by extracting iris feature from an iris image through the CPA based on the cumulative sum. A divider(310) converts the iris image into the images having a specific pixel size into polar coordinates and divides the polar coordinates into cells. A grouper(320) groups/outputs the first group comprising the fixed number of cells and the second group comprising the bigger number of cells than the first group. A feature vector generator(330) performs the cumulative sum-based CPA by using a feature value of each cell grouped in the first or second group as a representative value and generates a feature vector by assigning a different value to the cell corresponding to a change point and the remaining cells. An authenticator(340) performs authentication by comparing the generated feature vector with the registered feature vector.
Abstract translation: 提供了一种通过使用基于累积和的CPA来识别虹膜的方法和装置，以通过基于累积和通过CPA从虹膜图像中提取虹膜特征来对用户进行认证。 分频器（310）将虹膜图像转换为具有特定像素尺寸的图像成极坐标，并将极坐标分割成单元。 石斑鱼（320）对包含固定数量的细胞的第一组进行分组/输出，第二组包括比第一组更多的细胞数。 特征向量生成器（330）通过使用分组在第一组或第二组中的每个单元的特征值作为代表值来执行基于累积和的CPA，并且通过向对应于变化的单元分配不同的值来生成特征向量 点和剩余的单元格。 认证器（340）通过将生成的特征向量与登记的特征向量进行比较来进行认证。
-
公开(公告)号：KR1020040055511A
公开(公告)日：2004-06-26
申请号：KR1020020082204
申请日：2002-12-21
Applicant: 한국전자통신연구원
IPC: G06F15/00
Abstract: PURPOSE: A method for managing a security attribute of system resources is provided to quickly obtain/set security information and efficiently use a security information file in order to manage the security information of a kernel level. CONSTITUTION: At the initialization of a system, the contents of a security information file are loaded to a preset area on a memory. The security information is stored by being indexed to the memory connection list(S304). In case that a security information obtaining request is present in the kernel level, the corresponding security information is accessed by searching the memory connection list(S306).
Abstract translation: 目的：提供一种用于管理系统资源的安全属性的方法，以快速获取/设置安全信息，并有效地使用安全信息文件，以便管理内核级别的安全信息。 规定：在系统初始化时，将安全信息文件的内容加载到内存中的预设区域。 通过索引到存储器连接列表来存储安全信息（S304）。 在核心级别中存在安全信息获取请求的情况下，通过搜索存储器连接列表来访问对应的安全信息（S306）。
-
公开(公告)号：KR1020030082187A
公开(公告)日：2003-10-22
申请号：KR1020020020800
申请日：2002-04-17
Applicant: 한국전자통신연구원
IPC: G06F12/14
CPC classification number: G06F21/6209 , G06F21/6218
Abstract: PURPOSE: A method for controlling access using a token having security attribute on a computer system is provided to keep security even if a storing device is flown out by storing a file after encryption using an encryption attribute, and to put no special limitation for conventional operation while keeping the security by making a system manager read only the encrypted contents for backup. CONSTITUTION: The computer system gives the token having the security attribute to a user permitted to access to a file(S101). A user process checks an access request for the file. It is judged that the access requested file has the token having the security attribute(S103). If the file has the token, the access to the file is permitted. If not, the access is permitted or refused by judging that the user requesting the access has the same token(S104).
Abstract translation: 目的：提供一种使用计算机系统上具有安全属性的令牌来控制访问的方法，以便即使通过使用加密属性在加密之后存储文件来存储存储设备，也保持安全性，并且对常规操作没有特别限制 同时通过使系统管理员只读取加密的内容进行备份来保持安全性。 构成：计算机系统向允许访问文件的用户给出具有安全属性的令牌（S101）。 用户进程检查文件的访问请求。 判断访问请求文件具有具有安全属性的令牌（S103）。 如果文件有令牌，则允许访问该文件。 如果不是，则通过判断请求访问的用户具有相同的令牌来允许或拒绝该访问（S104）。
-
公开(公告)号：KR1020030054657A
公开(公告)日：2003-07-02
申请号：KR1020010084866
申请日：2001-12-26
Applicant: 한국전자통신연구원
IPC: G06F15/00
Abstract: PURPOSE: A role based access control method is provided to control an access request based on a user role at a kernel level for preventing or intercepting a system hacking. CONSTITUTION: The method comprises several steps. A process selects effective roles among current roles by checking whether the roles, whose current member is the process, exist with the data on all the generated roles stored at a role data file(S1). The generation number of the effective roles, stored at the role data file, is compared with that of the roles, stored at an object security database, and a permission value of the role defined at an object is regarded as meaningless in a case that the two generation numbers are not same(S2). An OR operation is applied to attribute values defined at the object corresponding to the remaining roles processed via the steps, S1 and S2, it is checked whether a requested attribute value exists at the object, and an access permission result is output according to the check result(S3).
Abstract translation: 目的：提供一种基于角色的访问控制方法，以根据内核级别的用户角色来控制访问请求，以防止或拦截系统黑客。 构成：该方法包括几个步骤。 过程通过检查当前成员是进程的角色是否存在于角色数据文件（S1）中存储的所有生成的角色上的数据，来选择当前角色中的有效角色。 存储在角色数据文件中的有效角色的生成次数与存储在对象安全数据库中的角色的生成次数进行比较，并且在对象的定义角色的权限值被认为是无意义的， 两代数不相同（S2）。 对与通过步骤S1和S2处理的剩余角色相对应的对象定义的属性值应用OR操作，检查对象是否存在请求的属性值，并且根据检查输出访问许可结果 结果（S3）。
-
公开(公告)号：KR1020160142460A
公开(公告)日：2016-12-13
申请号：KR1020150078025
申请日：2015-06-02
Applicant: 한국전자통신연구원
CPC classification number: G06K9/4647 , G06K9/6292
Abstract: 본발명의일 실시예에따른객체인식장치는대상영상을참조하여복수의레벨영상을생성하는레벨영상생성부, 각레벨영상으로부터특징벡터를추출하는특징벡터추출부, 레벨영상별 특징벡터를클러스터링하여코드워드를생성하는코드워드생성부, 코드워드에상응하는히스토그램을생성하는히스토그램생성부및 히스토그램에따라대상영상에대한객체인식정보를생성하는분류기를포함한다.
Abstract translation: 根据示例的用于对象检测的装置包括：水平图像生成单元，被配置为参考目标图像生成多个级别图像; 特征向量提取单元，被配置为从每个级别图像提取特征向量; 码字生成单元，被配置为通过对每个级别图像的特征向量进行聚类来生成码字; 直方图生成单元，被配置为生成与所述码字对应的直方图; 以及分类器，被配置为基于直方图生成目标图像的对象识别信息。
-
-
-
-
-
-
-
-
-
Search Results
63
records
(took 0.027 seconds)
