자연어를 이용한 이미지 데이터 색상 및 형태 검색 장치및 그 방법
    61.
    发明授权
    자연어를 이용한 이미지 데이터 색상 및 형태 검색 장치및 그 방법 失效
    자연어를이용한이미지데이터색상및형태검색장치및그방

    公开(公告)号:KR100422805B1

    公开(公告)日:2004-03-12

    申请号:KR1020000082179

    申请日:2000-12-26

    CPC classification number: G06F17/30259 G06F17/3025

    Abstract: An apparatus and method for searching color and shape of image data based on a natural language with Fuzzy concept is disclosed. The apparatus for retrieving includes: a dictionary storing unit for storing a dictionary used for processing a natural language; a color/shape threshold storing unit for storing color histograms mapped to color related words and edge information corresponding to shape related words; a query input unit for receiving a query sentence which describes the color and the shape of the image by using a natural language; an analyzing unit for analyzing the query sentence based on the dictionary information and generating analyzed words; a color/shape recognizing unit for recognizing whether the analyzed words represent the color or the shape; a color/shape threshold database constructing unit for mapping and storing color histograms to color related words and storing edge information corresponding to shape related words; a color/shape threshold retrieving unit for retrieving the color histograms and the edge information corresponding to the analyzed words from the color/shape threshold storing unit; and a retrieving result output unit for providing image data searched in the color/shape threshold retrieving unit.

    Abstract translation: 公开了一种用于使用Fuzzy概念基于自然语言来搜索图像数据的颜色和形状的装置和方法。 该检索设备包括:字典存储单元,用于存储用于处理自然语言的字典; 颜色/形状阈值存储单元,用于存储映射到颜色相关字的颜色直方图和与形状相关字对应的边缘信息; 查询输入单元,用于通过使用自然语言接收描述图像的颜色和形状的查询语句; 分析单元,用于基于词典信息分析查询语句并生成分析的词; 颜色/形状识别单元,用于识别分析的单词是代表颜色还是形状; 颜色/形状阈值数据库构建单元,用于将颜色直方图映射和存储到颜色相关词并存储对应于形状相关词的边缘信息; 颜色/形状阈值检索单元,用于从颜色/形状阈值存储单元检索与分析的单词对应的颜色直方图和边缘信息; 以及检索结果输出单元,用于提供在颜色/形状阈值检索单元中搜索到的图像数据。

    유,알,엘 포함관계에 기반한 유사도 재계산을 통한효과적인 홈페이지 검색 방법
    62.
    发明公开
    유,알,엘 포함관계에 기반한 유사도 재계산을 통한효과적인 홈페이지 검색 방법 失效
    通过网址重建来有效搜索主页的方法,包括基于关系的相似性

    公开(公告)号:KR1020030080826A

    公开(公告)日:2003-10-17

    申请号:KR1020020019647

    申请日:2002-04-11

    CPC classification number: G06F17/30864

    Abstract: PURPOSE: A method for effectively searching a homepage through the recalculation of the URL(Uniform Resource Locator) included relation based similarity is provided to easily know whether a searched web document is the homepage or not without visiting the URL of the searched web document. CONSTITUTION: A text is extracted from the searched web document according to a query from a user(S400). An index file for performing the web searching is generated by indexing the extracted text(S406). The web document and a searching result ranking the web document by considering the weight of a searching query are output. The similarity for the ranked searching result web documents is recalculated by using the URL included relation of the searched web document. The web document corresponding to the homepage is displayed with a high priority by rearranging the rank of the web document through the recalculated similarity.

    Abstract translation: 目的:提供一种通过重新计算URL(统一资源定位器)包括基于关联的相似度来有效搜索主页的方法,以便在不访问所搜索的网络文档的URL的情况下容易地知道所搜索的网页文档是否是主页。 构成:根据来自用户的查询从所搜索的web文档中提取文本(S400)。 通过索引提取的文本来生成用于执行网页搜索的索引文件(S406)。 输出Web文档和通过考虑搜索查询的权重对Web文档排序的搜索结果。 通过使用所搜索的web文档的URL包括的关系来重新计算排名的搜索结果web文档的相似性。 通过重新计算相似度来重新排列Web文档的等级,通过高优先级显示对应于主页的web文档。

    문서분류기법을 이용한 정답문서집합 자동 구축 방법
    63.
    发明公开
    문서분류기법을 이용한 정답문서집합 자동 구축 방법 有权
    使用文件分类方法自动构建正确答案文件的方法

    公开(公告)号:KR1020030056541A

    公开(公告)日:2003-07-04

    申请号:KR1020010086769

    申请日:2001-12-28

    Abstract: PURPOSE: A method for automatically constructing a correct answer document set using a document classification method is provided to automatically construct a correct document of the information searching and the Internet answer-response for the vast Internet documents through a meaning analysis of a natural language query by automatically classifying the correct answer document set. CONSTITUTION: The highly related document is extracted by collecting the document related to a concept word(S410). The attribute is defined into each theme by analyzing the collected document(S420). A learning document for the automatic classification for the defined attribute is constructed(S430). An automatic classifier is made by using the constructed learning document(S440). The concept word having the similar attribute is extracted from the concept network(S450). An attribute rule is extracted from the document allocated to the attribute(S460). The constructed document is automatically classified by using the extracted concept relationship and the attribute rule(S470). A new attribute rule is extracted from the automatically classified document(S480).

    Abstract translation: 目的:提供一种使用文档分类方法自动构建正确答案文档集的方法,通过自然语言查询的意义分析自动构建大型互联网文档的信息搜索和互联网应答响应的正确文档, 自动分类正确的答案文件集。 构成:通过收集与概念词相关的文档来提取高度相关的文档(S410)。 通过分析收集的文档将属性定义到每个主题中(S420)。 构建用于定义属性的自动分类的学习文档(S430)。 使用构建的学习文档制作自动分类器(S440)。 从概念网络提取具有相似属性的概念词(S450)。 从分配给属性的文档中提取属性规则(S460)。 通过使用提取的概念关系和属性规则自动分类构建的文档(S470)。 从自动分类的文档中提取新的属性规则(S480)。

    인터넷 질의/응답을 위한 지식베이스 구축방법
    64.
    发明公开
    인터넷 질의/응답을 위한 지식베이스 구축방법 有权
    用于构建互联网问题知识库的方法/解答

    公开(公告)号:KR1020020058533A

    公开(公告)日:2002-07-12

    申请号:KR1020000086642

    申请日:2000-12-30

    Abstract: PURPOSE: A method for constructing a knowledge base for an Internet question/answer is provided to interpret a meaning of a query language exactly and supply proper information according to a using area of concept languages and an attribute by extending a conventional concept network, thereby collecting and classifying web documents based on a definition of a knowledge base and a concept language existed in a concept network. CONSTITUTION: A concept network is constructed by collecting concept languages for constructing a knowledge base(301). A web page search with respect to a constructed concept language is performed using a web page search, and candidates of a correct answer document are collected(302). Related documents are extracted and classified with respect to the collected documents based on a pre-defined attribute(303). A representative attribute is defined and decided with respect to a document group not embodied by an attribute(304). If attributes with respect to each document group are defined, similar attributes are integrated(305). An integration of attributes is achieved to the upper concept(306).

    Abstract translation: 目的:提供一种用于构建互联网问题/答案知识库的方法,以精确地解释查询语言的含义,并通过扩展常规概念网络,根据概念语言和属性的使用区域提供适当的信息,从而收集 并且基于概念网络中存在的知识库和概念语言的定义对web文档进行分类。 构成:通过收集构建知识库的概念语言构建概念网络(301)。 使用网页搜索来执行关于构造的概念语言的网页搜索,并且收集正确答案文档的候选(302)。 基于预定义属性(303),相对于所收集的文档提取和分类相关文档。 关于不由属性(304)体现的文档组来定义和确定代表属性(304)。 如果定义相对于每个文档组的属性,则集成类似的属性(305)。 对上层概念实现了属性的集成(306)。

    키팩트기반 텍스트검색모델의 검색성능측정방법
    65.
    发明公开
    키팩트기반 텍스트검색모델의 검색성능측정방법 失效
    基于事实的文本检索模型的检索性能测量方法

    公开(公告)号:KR1020010001989A

    公开(公告)日:2001-01-05

    申请号:KR1019990021549

    申请日:1999-06-10

    Abstract: PURPOSE: A retrieval performance measurement method for a key fact based text retrieval model is provided to calculate a weight value on a document and a query language and give the document a priority based on a similarity between the document and the query. CONSTITUTION: A retrieval performance measurement method for a key fact based text retrieval model comprises the steps of extracting key words from a set of documents and queries(S01), asking if setting a variable value of retrieval performance assessment elements on the extracted result(S02), if the answer is positive, giving the variable value of retrieval performance assessment elements on a new document size, a key fact frequency within the document, and a key fact weight constant(S03), making a document and a query vector including the assessment variable value(S04), calculating a key fact weight value of the document and the query by a key fact weight equation(S05), obtaining a retrieval performance result of the key fact based text retrieval model for searching for an optimal retrieval model according to characteristics of a retrieval target document domain(S06), giving the documents a priority number(S07), and assessing the retrieval model(S08).

    Abstract translation: 目的:提供一种基于关键事实的文本检索模型的检索性能测量方法,以计算文档和查询语言的权重值,并根据文档和查询之间的相似性给出文档的优先级。 构成:用于基于关键事实的文本检索模型的检索性能测量方法包括从一组文档和查询中提取关键词的步骤(S01),询问是否在提取结果上设置检索性能评估要素的可变值(S02 ),如果答案是肯定的,则将新的文档大小,文档中的关键事实频率以及关键事实权重常数(S03)给出检索性能评估元素的变量值,使得包括文档的查询向量 评估变量值(S04),通过关键事实权重方程(S05)计算文档和查询的关键事实权重值(S05),获得基于关键事实的文本检索模型的检索性能结果,以搜索最佳检索模型 检索目标文档域的特征(S06),给出文档优先权数(S07),并评估检索模型(S08)。

    용언의 문형정보를 이용한 선택 제약 사전 구축 방법 및 시스템
    67.
    发明授权
    용언의 문형정보를 이용한 선택 제약 사전 구축 방법 및 시스템 有权
    使用预测模式构建选择限制字典的方法和系统

    公开(公告)号:KR101709185B1

    公开(公告)日:2017-02-23

    申请号:KR1020140162399

    申请日:2014-11-20

    Abstract: 본발명은용언의문형정보를이용한선택제약사전구축방법및 시스템을제공한다. 상기용언의문형정보를이용한선택제약사전구축방법은용언의문형정보를획득하는단계, 상기획득한용언의문형정보에서용언을추출하고, 상기추출된용언이포함된용례문장을검색하는단계, 상기검색된용례문장의의미를분석하여상기용언과다른의미의용언이포함된용례문장을필터링하는단계, 상기필터링된용례문장의구문을분석하여상기용언과관련없는구성성분을필터링하는단계, 상기필터링된용례문장에서선택제약대상을선택하고, 어휘의미망을이용하여상기선택된선택제약대상의상위경로를추출하는단계및 상기추출된상위경로들의비교를통해상기용언에대한선택제약경로를설정하고, 상기설정된선택제약경로를이용하여상기용언에대한선택제약사전을구축하는단계를포함한다.

    Abstract translation: 本发明提供一种使用谓词的句子模式构建选择限制词典的方法,其能够通过使用与句子的谓词相关的信息来检测句子的结构和含义。 使用谓词的句型构建选择限制词典的方法包括以下步骤:获取谓词的句型; 从谓词的获取句型中提取谓词以搜索包括提取的谓词的句型; 分析句子模式的含义以过滤包括具有与谓词不同的含义的谓词的句子模式; 分析过滤句子模式的结构,以过滤与谓词无关的组件; 从所述过滤的句子中选择所述选择限制的目标,以通过使用词汇sematic网络来提取所述选择限制的所述选定目标的高级路径; 并且通过将所提取的高级路径彼此进行比较来设置谓词的选择限制路径,以及为谓词构建选择限制字典。

    언어분석결과 업데이트 장치 및 방법
    69.
    发明公开
    언어분석결과 업데이트 장치 및 방법 审中-实审
    装置和方法校正语言分析结果

    公开(公告)号:KR1020160060820A

    公开(公告)日:2016-05-31

    申请号:KR1020140162397

    申请日:2014-11-20

    CPC classification number: G06F17/27

    Abstract: 본발명은대용량언어분석결과중 부정확한부분을자동탐색하여업데이트하는언어분석결과업데이트장치및 방법에관한것으로서, 본발명에따른언어분석결과업데이트장치는언어분석결과및 상기언어분석결과의업데이트에사용될언어분석메타데이터를저장하는저장부; 및언어지식리소스에추가된언어지식을기반으로상기언어분석메타데이터를재분석하고, 재분석결과를기반으로상기언어분석결과를업데이트하는업데이트부를포함한다.

    Abstract translation: 本发明涉及一种用于更新语言分析结果的装置和方法,其自动探索批量语言分析结果的不准确部分并更新它们。 根据本发明,用于更新语言分析结果的装置包括:存储单元,其存储用于更新语言分析结果的语言分析结果和语言分析元数据; 以及更新单元,其基于添加到语言知识资源的语言知识重新分析语言分析元数据,并且基于重新分析结果来更新语言分析结果,从而实时地提高语言分析结果。

    자연어 질의 응답 시스템 및 방법
    70.
    发明公开
    자연어 질의 응답 시스템 및 방법 审中-实审
    自然语言问答系统和方法

    公开(公告)号:KR1020160060253A

    公开(公告)日:2016-05-30

    申请号:KR1020140161904

    申请日:2014-11-19

    Abstract: 본발명은자연어질의응답시스템및 방법에관한것으로, 상기자연어질의응답시스템은, 저장부에저장되어있는비정형문서의텍스트를분석하여, 문장이나단락의의미에따라비정형문서를분류및 색인하는색인부; 상기색인부로부터전송되는의미별로색인된비정형문서를수신및 저장하는데이터베이스; 입력되는질의어를의미적으로분석하여색인어를추출하고, 추출된색인어와관련된문서를상기데이터베이스에서검색하는검색부; 및상기검색부에의해검색된문서를분석하여, 질의어에대한정답을생성하고, 검색결과와정답을제공하는제공부로구성될수 있다.

    Abstract translation: 本发明涉及一种自然语言查询 - 响应系统及其方法。 该系统包括:分析存储在存储单元中的非结构化文档的文本的索引单元,并且通过句子或段落的含义对非结构化文档进行分类和索引; 数据库,其接收并存储由所述含义索引的所述非结构化文档,其中,从所述索引单元发送所述非结构化文档; 搜索单元,通过语义分析输入的查询和搜索数据库中提取与提取的索引词相关的文档来提取索引词; 以及提供单元,分析由搜索单元搜索到的文档,生成查询的答案,以及提供搜索结果和答案。 本发明提供了一种自然语言查询 - 响应系统,其被配置为改善信息搜索性能并提供正确答案。

Patent Agency Ranking