-
公开(公告)号:KR101366636B1
公开(公告)日:2014-03-13
申请号:KR1020120091616
申请日:2012-08-22
Applicant: 주식회사신도리코 , 서울대학교산학협력단
IPC: H04N1/00
CPC classification number: H04N1/00005 , H04N1/00007 , H04N1/00031
Abstract: The present invention relates to a device and a method for evaluating the quality of a printed image. The device for evaluating the quality of a printed image according to the present invention includes: a binary coding unit converting an evaluation image having a dot area into a binary coded image; a grid extracting unit extracting a grid for separating the dot areas from the binary coded image converted in the binary coding unit; and an image evaluation unit evaluating the binary coded image using the dot area in each grid. Therefore, the concentration of a printed image can be evaluated. [Reference numerals] (110) Evaluation image input unit; (120) Binary coding unit;; (130) Dot area center extracting unit; (140) Grid extracting unit; (150) Valid area selection unit; (160) Image evaluation unit
Abstract translation: 本发明涉及一种用于评估打印图像的质量的装置和方法。 根据本发明的用于评估打印图像的质量的装置包括:二进制编码单元,将具有点区域的评估图像转换为二进制编码图像; 网格提取单元,从二进制编码单元中转换的二进制编码图像提取用于分离点区域的网格; 以及图像评估单元,使用每个网格中的点区域评估二进制编码图像。 因此,可以评估印刷图像的浓度。 (附图标记)(110)评价图像输入单元; (120)二进制编码单位; (130)点区中心提取单元; (140)网格提取单元; (150)有效区域选择单位; (160)图像评估单元
-
公开(公告)号:KR1020120103284A
公开(公告)日:2012-09-19
申请号:KR1020110021444
申请日:2011-03-10
Applicant: 서울대학교산학협력단
CPC classification number: G06K9/6267 , G06F17/30244 , G06K9/4671
Abstract: PURPOSE: An image processing method and an apparatus thereof through a repetition pattern are provided to efficiently match an image including a plurality of repetition patterns and to efficiently recognize an image including a repetition pattern. CONSTITUTION: A feature extracting unit(120) extracts a plurality of feature points from an inputted image. A classifying unit(130) classifies a plurality of feature points into one or more first feature points and one or more second feature points extracted from a repeated pattern. The classifying unit generates a classifier about each repetition pattern by using the first feature points. [Reference numerals] (110) Inputting unit; (120) Feature extracting unit; (130) Classifying unit; (140) Matching unit; (150) Image DB
Abstract translation: 目的:提供一种通过重复图案的图像处理方法及其装置,以有效地匹配包括多个重复图案的图像,并且有效地识别包括重复图案的图像。 构成:特征提取单元(120)从输入图像中提取多个特征点。 分类单元(130)将多个特征点分类为从重复模式提取的一个或多个第一特征点和一个或多个第二特征点。 分类单元通过使用第一特征点来生成关于每个重复图案的分类器。 (附图标记)(110)输入单元; (120)特征提取单元; (130)分类单位; (140)配套单元; (150)图像DB
-
公开(公告)号:KR101247220B1
公开(公告)日:2013-03-25
申请号:KR1020110021444
申请日:2011-03-10
Applicant: 서울대학교산학협력단
Abstract: 개시된 기술에 따른 영상 처리 방법은 (a) 입력된 이미지로부터 복수의 특징점들을 추출하는 단계, (b) 상기 복수의 특징점들을 반복적인 패턴들로부터 추출된 적어도 하나의 제1 특징점과 두드러진 적어도 일부로부터 추출된 적어도 하나의 제2 특징점으로 분류하는 단계 및 (c) 상기 제1 특징점들을 이용하여 상기 반복적인 패턴들 각각에 대한 분류기를 생성하는 단계를 포함한다.
-
公开(公告)号:KR101255841B1
公开(公告)日:2013-04-23
申请号:KR1020110001347
申请日:2011-01-06
Applicant: 서울대학교산학협력단
IPC: G06F17/30
Abstract: 이미지 연관 검색 방법은 제1 이미지 데이터베이스 및 제2 이미지 데이터베이스를 기초로 이루어진다. 상기 이미지 연관 검색 방법은 (a) 상기 제1 및 제2 이미지 데이터베이스에 포함된 이미지들을 상기 시각단어들의 빈도로 표현하는 단계, (b) 상기 시각단어들의 빈도를 기초로 바이소스 토픽모델을 생성하는 단계 및 (c) 상기 바이소스 토픽모델을 이용하여 상기 제1 이미지 데이터베이스와 상기 제2 이미지 데이터베이스를 연관하여 이미지를 검색하는 단계를 포함한다.
-
公开(公告)号:KR1020110085728A
公开(公告)日:2011-07-27
申请号:KR1020100005661
申请日:2010-01-21
Applicant: 삼성전자주식회사 , 서울대학교산학협력단
IPC: H04B1/40
CPC classification number: G06K9/00704 , G06K9/00671 , G06K9/4676 , H04B1/40 , G06T7/00
Abstract: PURPOSE: An apparatus and method for recognizing a building area in a portable terminal are provided to prevent building area recognition failure due to correspondent relation expectation failure of a specific point. CONSTITUTION: A mobile terminal selects an arbitrary reference point among the extracted feature point(301). The portable terminal compares a reference point with a feature point. The portable terminal includes similarity in the same group(307). The portable terminal confirms whether the group process is completed about the feature point(309).
Abstract translation: 目的:提供一种用于识别便携式终端中的建筑物区域的装置和方法,以防止由于特定点的对应关系期望失败导致的建筑物区域识别失败。 构成:移动终端在提取的特征点(301)中选择任意的基准点。 便携式终端将参考点与特征点进行比较。 便携式终端包括相同组中的相似性(307)。 便携式终端确认关于特征点的组处理是否完成(309)。
-
公开(公告)号:KR1020120079968A
公开(公告)日:2012-07-16
申请号:KR1020110001347
申请日:2011-01-06
Applicant: 서울대학교산학협력단
IPC: G06F17/30
CPC classification number: G06F17/30253 , G06F17/18
Abstract: PURPOSE: An associative image searching method based on a bi-source topic model and a system for the same are provided to calculate similarity based on topic distribution by expressing the association of data with the topic distribution. CONSTITUTION: Images included in first image database and second image database are expressed based on the frequency of visual words(S310). A bi-source topic model is generated based on the frequency of the visual words(S320). The first image database and the second image database are associated based on the bi-source topic model(S330). Images are searched based on the association of the first image database and the second image database.
Abstract translation: 目的:提供一种基于双源主题模型及其系统的关联图像搜索方法,通过表达数据与主题分布的关联,基于主题分布计算相似度。 构成:基于视觉词的频率来表示包括在第一图像数据库和第二图像数据库中的图像(S310)。 基于视觉词的频率生成双源主题模型(S320)。 第一图像数据库和第二图像数据库基于双源主题模型相关联(S330)。 基于第一图像数据库和第二图像数据库的关联来搜索图像。
-
-
-
-
-