-
公开(公告)号：KR101150754B1
公开(公告)日：2012-06-14
申请号：KR1020110032655
申请日：2011-04-08
Applicant: 계명대학교 산학협력단 , (주)제이브이엠
Abstract: PURPOSE: A drug box area detecting system and a method thereof are provided to easily confirm a loaded box state and to control the box state. CONSTITUTION: An image information extracting unit(31) detects the outline of a box. The image information extracting unit extracts the image information. An image analyzing unit(33) calculates the image information. A box control unit(35) generates the control signal according to a state of the box. The box control unit operates a loading box.
Abstract translation: 目的：提供一种药盒区域检测系统及其方法，以容易地确认装箱状态并控制盒状态。 构成：图像信息提取单元（31）检测框的轮廓。 图像信息提取单元提取图像信息。 图像分析单元（33）计算图像信息。 盒控制单元（35）根据盒子的状态产生控制信号。 箱控制单元操作装载箱。
-
公开(公告)号：KR1020100116404A
公开(公告)日：2010-11-01
申请号：KR1020090035088
申请日：2009-04-22
Applicant: 계명대학교 산학협력단
CPC classification number: G06K9/00147 , G06T5/20 , G06T7/11 , G06T7/149 , G06T7/33 , G06T2207/20116
Abstract: PURPOSE: A method and a device for classifying an independent cell and a group cell from image information are provided to classify cell areas into an independent cell area and a group cell area, thereby increasing reliability for cell diagnosis. CONSTITUTION: Image information is inputted and divided into a plurality of sub areas(204). Each sub area is binarized using a threshold value which is adaptively set according to distribution of cell image information for each sub area so that image information is divided into cell areas and a background area(206). The first feature vector showing a round degree of a cell shape for each cell area, the second feature vector showing a fixed distance from the center of a cell to an edge, and the third feature vector showing a shape of a cell are obtained.
Abstract translation: 目的：提供一种从图像信息中分离独立单元和组单元的方法和设备，用于将单元区分成独立的单元区域和组单元区域，从而增加了单元诊断的可靠性。 构成：图像信息被输入并分成多个子区域（204）。 使用根据每个子区域的小区图像信息的分布自适应地设置的阈值来对每个子区域进行二值化，使得图像信息被划分为小区区域和背景区域（206）。 显示每个单元格区域的单元形状的圆度的第一特征向量，示出从单元格的中心到边缘的固定距离的第二特征向量和表示单元格形状的第三特征向量。
-
公开(公告)号：KR1020090106133A
公开(公告)日：2009-10-08
申请号：KR1020080031653
申请日：2008-04-04
Applicant: 계명대학교 산학협력단
IPC: H04N7/18
CPC classification number: G08B17/125 , G06K9/6296 , G06T7/277
Abstract: PURPOSE: A fire detecting device using Bayesian network and a method thereof using a probability model are provided to improve detection performance and reduce calculation time by applying probability model as value for Bayesian inference. CONSTITUTION: A fire detection method is as follows. A moving area is detected by inputting video data. The pixels of fire color are extracted in video data according to fire color probability mode by inputting the video data in a fire pixel detection module(104). High frequency components such as color value, a horizontal line, a vertical line and a diagonal line are detected about pixels of the fire color within the moving area. A first value which is combine probability between high frequency elements is calculated.
Abstract translation: 目的：提供一种使用贝叶斯网络的火灾探测装置及其使用概率模型的方法，通过应用概率模型作为贝叶斯推理值，提高检测性能，缩短计算时间。 构成：火灾探测方法如下。 通过输入视频数据来检测移动区域。 通过在火焰像素检测模块（104）中输入视频数据，根据火色概率模式，在视频数据中提取火色像素。 检测关于移动区域内的火色的像素的高频分量，例如颜色值，水平线，垂直线和对角线。 计算高频元素之间的组合概率的第一值。
-
公开(公告)号：KR100918436B1
公开(公告)日：2009-09-24
申请号：KR1020070121221
申请日：2007-11-27
Applicant: 계명대학교 산학협력단 , (주)위니텍
Abstract: 본 발명은 영상 정보로부터 화재 영역을 감지하는 비전 기반 화재 감지 시스템 및 방법에 관한 것이다.
상기한 화재 감지 시스템은, 이전 프레임의 그레이 영상을 배경으로 초기화한 후에 현재 프레임을 반영하여 이전 배경 프레임을 업데이트하여 출력하는 움직임 감지모듈; 상기 움직임 감지모듈이 출력하는 프레임에 대해 컬러 모델을 적용하여 제1화재 후보 영역을 검출하는 컬러 모델링 모듈; 상기 현재 프레임에 대해 루미넌스 맵을 적용하여 주변에 비해 밝은 제2화재후보영역을 검출하는 루미넌스 맵 처리 모듈; 상기 제1 및 제2화재후보영역에 대해 SVM(Support Vector Machine)을 이용하여 시간적인 웨이블릿 변환 에너지의 변화량 모델에 따라 최종 화재 감지 영역을 추출하는 SVM 처리 모듈을 구비하는 것을 특징으로 한다.
화재 감지, 영상, 컬러 모델링, 움직임 감지, 루미넌스 맵, SVM(Support Vector Machine)-
公开(公告)号：KR1020090054522A
公开(公告)日：2009-06-01
申请号：KR1020070121221
申请日：2007-11-27
Applicant: 계명대학교 산학협력단 , (주)위니텍
CPC classification number: G08B17/125 , G06K9/00 , G06T7/20 , G06T7/40 , H04N7/18
Abstract: A fire detecting system and a detecting method thereof are provided to detect a fire region by analyzing an amount of movement change and color information. A movement detecting module(100) updates a previous background frame by reflecting a current frame to a gray image of the previous frame. A color modeling module(102) detects a first fire candidate region by applying a color model to the frame outputted from the movement detecting module. A luminance map process module(104) detects a second fire candidate region brighter than a peripheral region by applying a luminance map to the current frame. An SVM(Support Vector Machine) process module(106) extracts the final fire detection region according to the change quantity model of the timely wavelet conversion energy by using the SVM.
Abstract translation: 提供火灾检测系统及其检测方法，通过分析移动量变化和颜色信息来检测火灾区域。 移动检测模块（100）通过将当前帧反映到前一帧的灰度图像来更新先前的背景帧。 颜色建模模块（102）通过将颜色模型应用于从运动检测模块输出的帧来检测第一火灾候选区域。 亮度图处理模块（104）通过将亮度图应用于当前帧来检测比周边区域更亮的第二火灾候补区域。 SVM（Support Vector Machine，支持向量机）处理模块（106）通过使用SVM，根据及时小波变换能量的变化量模型提取最终的火灾探测区域。
-
Patent Agency Ranking
公开(公告)号：KR1020080021181A
公开(公告)日：2008-03-07
申请号：KR1020060081846
申请日：2006-08-28
Applicant: 계명대학교 산학협력단 , 서미숙 , 고병철 , 남재열 , 원치선
CPC classification number: G06T1/0007 , G06K9/48 , G06T5/10 , G06T5/50 , G06T7/12 , G06T2207/20064 , G06T2207/20096 , G06T2210/41
Abstract: An image data processing method and a system thereof are provided to shorten a processing time for segmentation for an attention window and extract an interested region according to the visual characteristics of the human beings. An AW(Attention Window) generating unit(102) receives image data and converts it into gray level degree image data to generate a normalized contrast map, and wavelet-converts the image data to extract an AW of the image data based on points with a variation greater than a certain value of a frequency and the contrast map. An AW segmentation unit(104) segments the AW through dividing using an average value of the contrast map and quadtree process. A weight value setting unit(106) assigns a weight value of each AW to each segmented AW by using a contrast comparison map matched to human visual characteristics. A feature vector generating unit(108) generates a feature vector with respect to the segmented AW.
Abstract translation: 提供一种图像数据处理方法及其系统，以缩短用于注意窗口的分割的处理时间，并根据人的视觉特征提取兴趣区域。 AW（注意窗口）生成单元（102）接收图像数据并将其转换为灰度级图像数据以生成归一化对比度图，并且对图像数据进行小波变换以提取图像数据的AW， 变化大于某一特定频率值和对比度图。 AW分割单元（104）使用对比度图和四叉树处理的平均值通过划分来划分AW。 权重值设定单元（106）通过使用与人类视觉特征相匹配的对比度对比图来为每个分割AW分配每个AW的权重值。 特征向量生成单元（108）相对于分割的AW生成特征向量。
-
公开(公告)号：KR1020080020257A
公开(公告)日：2008-03-05
申请号：KR1020060083515
申请日：2006-08-31
Applicant: 계명대학교 산학협력단 , 서미숙 , 고병철 , 남재열
CPC classification number: G06F17/30799 , G06T7/13
Abstract: A method and an apparatus for bio-image retrieval using the characteristic edge block of an edge histogram descriptor are provided to generate a characteristic vector using the characteristic edge block of an image and local dispersion and perform similarity comparison using the characteristic vector, thereby providing the location information of a main object existing in the image and improving search performance by reducing the error value of similarity comparison caused by the comparison of an edge block with no characteristic. A similarity comparison process for bio-image retrieval comprises the following steps of: collecting original bio-images and constructing a DB(Database) with bio-images generated by conversion processes(S501); storing the local and global edge histogram of reference images inputted to the DB in the DB(S502); generating a characteristic vector for an inquiry image(S503,S504); selecting the same block of the reference image existing in the DB for the selected characteristic edge block of the inquiry image to generate the characteristic vector of the reference image and accumulating the absolute error of two characteristic vector values of the inquiry image and the reference image(S505,S506); and arranging the reference images in an order in which the size of the sum of accumulated errors is arranged in a small size and providing parent reference images to a search system user as result(S507).
Abstract translation: 提供使用边缘直方图描述符的特征边缘块进行生物图像检索的方法和装置，以使用图像的特征边缘块和局部色散生成特征向量，并使用特征向量进行相似性比较，从而提供 存在于图像中的主要对象的位置信息，并且通过减少由没有特征的边缘块的比较引起的相似性比较的误差值来提高搜索性能。 用于生物图像检索的相似性比较过程包括以下步骤：收集原始生物图像并用转化过程产生的生物图像构建DB（数据库）（S501）; 将输入到DB的参考图像的局部和全局边缘直方图存储在DB中（S502）; 生成查询图像的特征向量（S503，S504）; 为查询图像的所选择的特征边缘块选择存在于DB中的参考图像的相同块，以生成参考图像的特征向量，并累加询问图像和参考图像的两个特征向量值的绝对误差（ S505，S506）; 并且按照将累积误差之和的大小排列成小尺寸并将结果提供给搜索系统用户的父参考图像的顺序排列参考图像（S507）。
-
公开(公告)号：KR1020060133610A
公开(公告)日：2006-12-27
申请号：KR1020050053294
申请日：2005-06-21
Applicant: 계명대학교 산학협력단
IPC: A61B7/04
Abstract: A heart sound classification method using a Hidden Markov Model(HMM) is provided to judge heart sound reliably, by providing an automatic heart sound classification method using HMM instead of ANN(Artificial Neural Network). In a heart sound classification method, a Hidden Markov Model(HMM) parameter value(40) as to heart sound data(10) is estimated. A sort of decease corresponding to the given heart sound data is determined by using the HMM parameter value. The HMM parameter estimation process includes a process of setting an HMM initial parameter value as to the heart sound data, and a process of re-estimating the HMM parameter value through an Expectation Maximization process of the HMM initial parameter value.
Abstract translation: 提供使用隐马尔可夫模型（HMM）的心音分类方法，通过使用HMM代替ANN（人造神经网络）提供自动心音分类方法来可靠地判断心音。 在心音分类方法中，估计关于心音数据（10）的隐马尔可夫模型（HMM）参数值（40）。 通过使用HMM参数值来确定对应于给定心音数据的一种死亡。 HMM参数估计处理包括设置关于心音数据的HMM初始参数值的处理，以及通过HMM初始参数值的期望最大化处理重新估计HMM参数值的处理。
-
公开(公告)号：KR1020170062661A
公开(公告)日：2017-06-08
申请号：KR1020150168074
申请日：2015-11-28
Applicant: 계명대학교 산학협력단
Abstract: 본발명은컬러영상에서 Mean-Shift 군집화와단계별병합방법을이용한자동원료선별방법에관한것으로서, 보다구체적으로는 (1) 컬러로촬영된원료영상을입력받는단계,(2) 상기입력받은컬러원료영상에서, 배경을제거하여전경맵영상을추출하는단계, (3) 상기추출된전경맵영상을 Mean-Shift 군집화알고리즘에적용하여상기전경맵영상을 N개의군집으로나누는단계, (4) 상기나누어진 N개의군집에서가장큰 군집을 seed 군집으로선택하는단계, (5) 상기선택된 seed 군집과나머지각 군집에대해, 상기 seed 군집과주변군집간의위치근접성, 및상기 seed 군집의대표색상과주변군집의대표색상간의색상유사성을측정하여, 상기위치근접성및 색상유사성이임계값이하일경우, 두군집들을병합하는단계, (6) 상기병합된군집데이터를 RG/GB/BR의 2차원컬러분포도로변환하여표현하고, 상기각 컬러분포도의군집데이터를기준으로타원을생성하는단계, 및 (7) 상기생성된타원에포함되는지여부를기준으로원료의양품과불량품을분류하는단계를포함하는것을그 구성상의특징으로한다. 본발명에서제안하고있는컬러영상에서 Mean-Shift 군집화와단계별병합방법을이용한자동원료선별및 장치에따르면, 컬러영상을기반으로 Mean-Shift 군집화와단계별병합방법을이용하여자동으로원료를선별함으로써, 다양한색상의원료에대해사용자의인위적조작이적고, 보다정확하게자동으로원료를선별할수 있다.
-
公开(公告)号：KR1020170033716A
公开(公告)日：2017-03-27
申请号：KR1020150131850
申请日：2015-09-17
Applicant: 계명대학교 산학협력단 , 이인텔리전스 주식회사
Abstract: 본발명은공간피라미드특징을이용한실시간속도제한표지판검출및 속도인식방법에관한것으로서, 보다구체적으로는 (1) 입력영상에서속도제한표지판후보영역을검출하는단계(S100); (2) 상기검출된후보영역이속도제한표지판에해당하는지여부를검증하는단계(S200); 및 (3) 상기검증된속도제한표지판의속도를인식하는단계(S300)를포함하는것을그 구성상의특징으로한다. 본발명에서제안하고있는공간피라미드특징을이용한실시간속도제한표지판검출및 속도인식방법과그에관한장치에따르면, 공간피라미드특징과 BoostRandomForest를속도제한표지판의검출및 검출된속도제한표지판의속도를인식하는데사용함으로써, 속도제한표지판을검출하고검출된속도제한표지판의속도를인식하는과정의속도를향상시켜준다. 또한, 본발명은, 실시간으로속도제한표지판을검출하고그에대한속도를빠르게인식함으로써, 운전자에게가장최신의교통정보를제공하여운전자의편의및 안전을증진시켜준다.
Abstract translation: 本发明包括用于检测在所述候选区中的限速标志，更具体地，步骤（S100）（1）与使用一个特征空间金字塔实时限速标志的检测和识别率方法的输入图像; （2）验证检测到的候选区域是否对应于限速标志（S200）; （3）识别验证的限速标志的速度（S300）。 根据实时限速标志检测及速度识别方法和装置，其使用本发明，空间金字塔特征提出并用于BoostRandomForest空间金字塔特征识别检测的速度和的限速标志检测的限速标志 从而加快检测限速标志的过程并识别检测到的限速标志的速度。 另外，本发明中，通过实时检测限速标志和快速地识别其速度，以提供最先进的最新的交通信息提供给司机有助于增加驱动器的方便性和安全性。
-
-
-
-
-
-
-
-
-
Search Results
73
records
(took 0.025 seconds)
