公开(公告)号：KR100472069B1

公开(公告)日：2005-03-10

申请号：KR1020020078287

申请日：2002-12-10

Applicant: 한국전자통신연구원

Inventor： 전병태 ,  이형구 ,  조수현 ,  정연구 ,  박선희

IPC: G06T1/00

Abstract: 본 발명은 세포 영상 분석 시스템에서의 세포 영상 분할 방법에 관한 것으로, 개선된 스네이크 영상 분할 방법을 이용하여 시간에 따라 변화하는 세포 영역을 효과적으로 분할 할 수 있는 세포 영상 분석 시스템에서의 세포 영상 분할 방법과 상기 방법을 실현시키기 위한 프로그램을 기록한 컴퓨터로 읽을 수 있는 기록매체를 제공하며, 입력된 전체영상에서 대표 영상을 추출하고, 각 세포의 경계선을 추출하여 첫 번째 영상의 스네이크 초기점으로 설정하는 제 1 단계; 상기 첫 번째 영상에서 세포의 경계선을 추출하여 다음 영상의 스네이크 초기점으로 설정하는 제 2 단계; 상기 제 2 단계에서 설정된 스네이크 초기점이 세포의 경계선 근처에 있는지 판단하는 제 3 단계; 상기 제 3 단계의 판단 결과, 세포의 경계선 근처에 있으면 유클리디언 거리 값으로 세포의 경계선을 추출하는 제 4 단계; 및 상기 제 3 단계의 판단 결과, 세포의 경계선 근처에 있지 않으면 블록 매칭을 수행하여 세포의 경계선까지 스네이크 점을 유도하고 상기 제 3 단계의 세포의 경계선 근처에 있는지 판단하는 과정으로 진행하는 제 5 단계를 포함하고, 바이오 영상 처리 시스템 등에 이용됨.

公开(公告)号：KR100449805B1

公开(公告)日：2004-09-22

申请号：KR1020010084869

申请日：2001-12-26

Applicant: 한국전자통신연구원

Inventor： 김계경 ,  정연구 ,  지수영 ,  황영섭 ,  유원필 ,  조수현 ,  이형구

IPC: G06K9/46

CPC classification number: G06K9/342 ,  G06K9/344 ,  G06K2209/01

Abstract: Disclosed is a method of segmenting touching numeral strings contained in handwritten touching numeral strings, and recognizing the numeral strings by use of feature information and recognized results provided by inherent structure of digits. The method comprises the steps of: receiving a handwritten numeral string extracted from a pattern document; smoothing a curved numeral image of the handwritten numeral string, and searching connecting components in the numeral image; determining whether or not the numeral string is a touching numeral string; if it is determined that the numeral string is the touching numeral string, searching a contour of the touching numeral string image; searching candidate segmentation points in the contour, and segmenting sub-images; computing a segmentation confidence value on each segmented sub-image by use of a segmentation error function to select the sub-image with the highest segmentation confidence value as a segmented numeral image in the touching numeral string image; if it is determined in the step c that the numeral string is not the touching numeral string, extracting a feature to recognize the segmented numeral image; segmenting the numeral image selected from the touching numeral string in the highest segmenting confidence value; and obtaining remaining numeral string image.

Abstract translation: 本发明公开了一种对包含在手写触摸数字串中的触摸数字串进行分段，并且通过使用特征信息和由数字的固有结构提供的识别结果来识别数字串的方法。 该方法包括以下步骤：接收从图案文档提取的手写数字串; 平滑手写数字串的曲线数字图像，并搜索数字图像中的连接分量; 确定数字串是否是触摸数字串; 如果确定数字串是触摸数字串，则搜索触摸数字串图像的轮廓; 搜索轮廓中的候选分割点，并分割子图像; 利用分割误差函数对每个分割子图像计算分割置信度值，以选择具有最高分割置信度值的子图像作为所述触摸数字串图像中的分割数值图像; 如果在步骤c中确定数字串不是触摸数字串，则提取识别分段数字图像的特征; 将从触摸数字串中选择的数字图像分割成最高分割置信度值; 并获得剩余的数字串图像。

公开(公告)号：KR1020040035389A

公开(公告)日：2004-04-29

申请号：KR1020020064519

申请日：2002-10-22

Applicant: 한국전자통신연구원

Inventor： 이재연 ,  정연구 ,  소정 ,  전병태 ,  윤호섭 ,  한규서 ,  김도형 ,  정용화

IPC: G06F15/00

CPC classification number: G06K9/00154

Abstract: PURPOSE: An online sign certification device using a pattern transform method and a realizing method thereof are provided to confirm an ID of a signer and enhance discrimination for a forged sign by analyzing a characteristic of a transform pattern visualizing a dynamic characteristic of a sign pattern. CONSTITUTION: A sign data input part(100) reads a locus of the sign as a sequence sampled in the same time interval by digitalizing the locus of the sign from a reference sign database(106). The first and the second pattern transformer(101,102) generate a transformed pattern sequence by respectively performing speed equalization and velocity transform for the read sign sequence. A characteristic extractor(104) generates three pairs of characteristic vectors by respectively extracting the characteristic vector from the read sign sequence and the pattern sequence transformed by the first/second pattern transformer. A difference vector calculator(108) generates a difference vector between the characteristic vector for a reference sign and the extracted characteristic vector. A judging part(110) judges accordance between the input sign and the reference by based on the difference vector.

Abstract translation: 目的：提供使用模式变换方法及其实现方法的在线标志认证装置，以通过分析可视化符号图案的动态特性的变换图案的特性来确认签名者的ID并增强对伪造符号的识别。 构成：符号数据输入部（100）通过对来自参照符号数据库（106）的符号的轨迹进行数字化，以同一时间间隔取样的顺序读取符号的轨迹。 第一和第二图案变换器（101,102）通过分别对读取符号序列执行速度均衡和速度变换来生成变换图案序列。 特征提取器（104）通过分别从读符号序列中提取特征向量和由第一/第二模式变换器变换的模式序列来产生三对特征向量。 差矢量计算器（108）产生用于参考符号的特征向量与提取的特征向量之间的差矢量。 判断部分（110）根据差矢量来判断输入符号和参考值之间的一致性。

公开(公告)号：KR100394276B1

公开(公告)日：2003-08-09

申请号：KR1019990028021

申请日：1999-07-12

Applicant: 한국전자통신연구원

Inventor： 구본기 ,  최이배 ,  정연구

IPC: G06T1/00

Abstract: PURPOSE: A method and a device for setting an initial hand area using a stereo interface method in a dactylology recognition are provided to automatically find an initial position of hands without additional actions or a specific equipment. CONSTITUTION: A hardware driver is driven to initialize cameras and a stereo stand. Images are obtained through two initialized cameras. Specific data are sampled from the stereo images by receiving a threshold value. A specific map is generated based on the specific data. Positive/negative area data according to the specific map are stored. Interference reference information is composed by the specific data. An interference window is generated by the interference reference information. A same point is searched from the correlation of two images by using the interface window and the specific map. Three-dimensional information is sampled. A variation map is sampled from the three-dimensional information. Distance information is stored from the variation map. The images are made of two color areas. An area of dactylology is searched.

Abstract translation: 目的：提供一种用于在指纹识别中使用立体界面方法来设置初始手区域的方法和设备，以在没有额外动作或特定设备的情况下自动找到手的初始位置。 构成：驱动硬件驱动程序初始化相机和立体声支架。 图像通过两个初始化的相机获取。 通过接收阈值从立体图像中采样特定数据。 根据特定数据生成特定映射。 存储根据特定地图的正面/负面区域数据。 干扰参考信息由特定数据组成。 干扰参考信息产生干扰窗口。 通过使用界面窗口和特定地图从两个图像的相关性中搜索相同的点。 三维信息被采样。 从三维信息中采样变化图。 距离信息是从变差图中存储的。 图像由两个颜色区域组成。 搜索指纹区域。

公开(公告)号：KR1020030054660A

公开(公告)日：2003-07-02

申请号：KR1020010084869

申请日：2001-12-26

Applicant: 한국전자통신연구원

Inventor： 김계경 ,  정연구 ,  지수영 ,  황영섭 ,  유원필 ,  조수현 ,  이형구

IPC: G06K9/46

CPC classification number: G06K9/342 ,  G06K9/344 ,  G06K2209/01

Abstract: PURPOSE: A written number sequence recognition method is provided to split the closely written numbers by using unique structural feature data of numbers and a recognition result, and to recognize each written number. CONSTITUTION: The method comprises several steps. A number recognition system receives a sequence of closely written numbers extracted from a document made with various written types of users(S1). The number recognition system smooths an indented number image(S2). The system searches for connection elements in order to classify the number types(S3). The system checks whether the number sequence is closely written(S4). In a case that the number sequence is closely written(S5), the system extracts a profile for the closely written numbers from the number image(S6). Then, the system splits the profile into assistant images by using candidate split points(S7), and calculates a split reliability value for each assistant image by using a defined split error function(S8). The system selects the assistant image of the highest split reliability value as a number image(S9). In a case that the number sequence is not closely written in the step S4 or the step S9 is performed, the system extracts feature data from the number image(S10), recognizes the corresponding number(S11), and in a case that there exists the next number image, the system repeats the step S4.

Abstract translation: 目的：提供一种书面序列识别方法，通过使用数字和识别结果的独特结构特征数据来分割紧密编写的数字，并识别每个书面编号。 构成：该方法包括几个步骤。 数字识别系统接收从由各种书写类型的用户制作的文档中提取的紧密编号的序列（S1）。 数字识别系统平滑缩进数字图像（S2）。 系统搜索连接元素以分类数字类型（S3）。 系统检查数字序列是否紧密写入（S4）。 在数字序列被紧密写入的情况下（S5），系统从数字图像中提取紧密写入的数字的简档（S6）。 然后，通过使用候选分割点（S7）将该轮廓分解为辅助图像，并且通过使用定义的分割误差函数来计算每个辅助图像的分割可靠度值（S8）。 系统选择最高分割可靠性值的辅助图像作为数字图像（S9）。 在步骤S4中没有紧密地写入数字序列或执行步骤S9的情况下，系统从数字图像中提取特征数据（S10），识别相应的数字（S11），并且在存在的情况下 下一个数字图像，系统重复步骤S4。

公开(公告)号：KR100369941B1

公开(公告)日：2003-01-30

申请号：KR1019990022152

申请日：1999-06-15

Applicant: 한국전자통신연구원

Inventor： 최이배 ,  정연구 ,  김승우 ,  오정택

IPC: G01B9/00

Abstract: PURPOSE: A method for compensating measurement error in phase-shifting project moire is provided to improve the accuracy of the three dimensional shape measurement of the whole system by compensating phase distortion error of the phase-shifting project moire. CONSTITUTION: A method for compensating measurement error in phase-shifting project moire includes the steps of obtaining a wavelength become longer than the original wavelength by generating beat phenomenon of the two wavelengths having similar characteristic value, calculating order of moire pattern by using the long wavelength generated by the beat, obtaining two equivalent wavelengths by using two grids having different pitch, extracting order of moire pattern phase about height of a measuring point(28) based on the short equivalent wavelength, and calculating the height of the measuring point from the extracted order.

Abstract translation: 目的：通过补偿移相项目莫尔相位失真误差，提出了一种补偿移相项目莫尔测量误差的方法，以提高整个系统三维形状测量的准确性。 本发明公开了一种补偿相移项目莫尔测量误差的方法，包括以下步骤：通过产生具有相似特征值的两个波长的跳动现象，获得波长变得比原始波长更长的波长，通过使用长波长 通过使用具有不同间距的两个网格获得两个等效波长，基于短等效波长提取关于测量点（28）的高度的莫尔图案相位的次序，并且从提取的图像中计算测量点的高度 订购。

公开(公告)号：KR100305368B1

公开(公告)日：2001-11-22

申请号：KR1019980045834

申请日：1998-10-29

Applicant: 한국정보화진흥원 ,  한국전자통신연구원

Inventor： 이원희 ,  황대준 ,  정연구 ,  서영호 ,  김용

IPC: G06Q50/10

Abstract: 본 발명은 멀티미디어 객체를 사용하는 화상회의와 상호참여 환경에서 사용되는 다양한 객체들 중 필요한 해당 객체만을 사용함으로써 사용자가 원하는 다양한 응용을 만들 수 있는 객체관리기 프레임워크의 객체생성 및 삭제방법에 관한 것으로서, 객체관리기를 이용하여 사용을 원하는 객체에 할당할 포트번호를 범역 세션 운영자(GSM)에게 요청하고 GSM으로 부터 제공된 포트번호를 객체관리기를 통하여 각 객체에게 전송하며, 이 값을 객체관리기가 세션에 참여중인 모든 참여자의 객체관리기에 전송하고, 세션관리기로부터 얻은 정보를 이용하여 각 객체간의 통신을 맺는 포트번호를 할당하고, 객체간의 통신을 종료시킨 후 상기 GSM이 다른 객체관리기에게 재 제공할 수 있도록 GSM에게 포트번호를 반환하는 포트번호 제거함으로써, 제한된 자원만으 로도 세션에 참여할 수 있게 되어 사용가용영역을 확장할 수 있고, 다양한 세션 형태를 만들어 갈 수 있으므로 세션의 지속성 및 다양성을 이룰 수 있으며, 다른 네트워크 프로그램과 발생할 지도 모를 오수신을 방지할 수 있는 효과가 있다.

公开(公告)号：KR100289048B1

公开(公告)日：2001-05-02

申请号：KR1019970069082

申请日：1997-12-16

Applicant: 한국전자통신연구원

Inventor： 김기홍 ,  정연구 ,  김용 ,  서영호

IPC: B62D65/00

Abstract: PURPOSE: System for automatically testing opening or closing of a window in a vehicle is provided to automatically alarm opening of the window. CONSTITUTION: System for automatically testing opening or closing of a window comprises a sensor(102) detecting wheels of a vehicle by being mounted on a front side of a moving direction of a conveyer belt(105); a racks(104) installed both sides of the conveyer belt; a CCD(Charge Coupled Device) camera(100) photographing window image of the vehicle while sensing the wheel of the vehicle by being mounted on the rack; and a PLC(Power Line Communication)(103) stopping the conveyer belt while opening the window. The CCD camera checks the opening or closing of the window. Thereby, the system for automatically testing opening or closing of the window in the vehicle automatically alarms the opening of the window through a warning lamp and a buzzer.

公开(公告)号：KR100269267B1

公开(公告)日：2000-10-16

申请号：KR1019970072162

申请日：1997-12-23

Applicant: 한국전자통신연구원 ,  한국과학기술원

Inventor： 정연구 ,  김용 ,  조강현 ,  김승우 ,  최이배

IPC: H04N13/00

Abstract: PURPOSE: A method for projection type moire using a phase transition grating and a system for the same are provided to improve measuring resolvability by applying a phase transition method to a projection type moire. CONSTITUTION: A projection grating(30) and a reference grating(31) are used as one couple. A relative distance between the projection grating(30) and the reference grating(31) is set up. Gratings of n couples are formed on one quartz glass by using a semiconductor lithography method. A phase transition grating(40) is manufactured by forming grating of n couples. The phase transition is obtained by using the phase transition grating(40).

Abstract translation: 目的：提供一种使用相变光栅的投射型莫尔条纹及其系统，以通过将相变法应用于投射型波纹来提高测量可分辨性。 构成：使用投影光栅（30）和参考光栅（31）作为一对。 设置投影光栅（30）和参考光栅（31）之间的相对距离。 通过使用半导体光刻方法，在一个石英玻璃上形成n个对的格栅。 通过形成n对的光栅来制造相变光栅（40）。 通过使用相变光栅（40）获得相变。