专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明专利

JPH07129563A METHOD FOR OBTAINING RELATIONAL WORD 失效
公开(公告)号：JPH07129563A
公开(公告)日：1995-05-19
申请号：JP26742993
申请日：1993-10-26
申请人： XEROX CORP
发明人： RONARUDO EMU KAPURAN , MAATEIN KEI
IPC分类号： G06F17/22 , G06F17/30
摘要： PURPOSE: To provide a method for speedily obtaining information related to the group of words such as synonyms. CONSTITUTION: A product is composed of a storage medium and word group data stored there and can be used for an electronic synonym dictionary or the like. The word group data contain the order strings of identifiers for words and grouping data expressing the group information concerning the group of identifiers. When the identifier coincident with the identifier of word is found out, while using the position and grouping data of the coincident identifier, data expressing the features of group including that word can be provided. For example, the group of identifiers can be ordered by length and the grouping data can express the numbers of groups for respective lengths. In that case, while using the position and grouping data of the coincident identifier, the position of identifiers in the groups are provided and they can be retrieved.

2. 发明专利

JPH0785047A PRODUCT WITH SET OF COMPACTLY ENCODED AND STORED STRINGS 失效
公开(公告)号：JPH0785047A
公开(公告)日：1995-03-31
申请号：JP19122993
申请日：1993-08-02
申请人： XEROX CORP
发明人： RONARUDO EMU KAPURAN , MAATEIN KEI , JIYON TEII MATSUKUSUUERU ZA SA
IPC分类号： G06F5/00 , G06F12/00 , G06F17/21 , G06F17/22 , G06F17/27 , G06F17/28
摘要： PURPOSE: To reduce the amount of data required for indicating a set of strings so as to compactly store the data for determining one set of the strings such as words and numerals. CONSTITUTION: An encoding process is started by receiving the main body of the data including the data relating to a state and the data relating to the corresponding output transition of the state as in a box 80. Then, in the box 82, the data relating to the respective states for deciding the order of relating transition units are obtained. In the box 84, the states are classified based on the number of input painters and the size of the pointer and the pointer or index of the size are allocated to the first output transition unit of the state provided with the input pointer whose subordinate block is larger than the pointer size. In the box 86, the transition unit is generated and a position is allocated to the transition unit. Then, in the box 88, a table and the transition unit are written in a file in the order based on the positions.

3. 发明专利

JPH076020A ENCODING SYSTEM OF FINITE STATE REGARDING HYPHENATION RULE 失效
公开(公告)号：JPH076020A
公开(公告)日：1995-01-10
申请号：JP7203594
申请日：1994-04-11
申请人： XEROX CORP
发明人： RONARUDO EMU KAPURAN , RAURI JIEI KARUTEYUNEN
IPC分类号： G06F5/00 , G06F17/24 , G06F40/191 , G06F17/26
摘要： PURPOSE: To provide a finite-state encoding system for hyphenation rule provided with a means which encodes a potentially infinite word list which is not required to be preknown in an FST structure for inspecting and generating hyphenation. CONSTITUTION: The proper position of a hyphen in an inputted string is decided by reading the symbol of the string through a finite-state converter 34 having a status-transition data structure decided by the compilation of a set of hyphenation rules and processing the symbol. The output of an encoding system can contain the indication of whether or not the output can contain strings connected with a hyphen or the strings coupled with a hyphen are accepted and whether or not the hyphenation inputted by the set of hyphenation rules is appropriate.

4. 发明专利

JPH06208453A METHOD FOR CONSTRUCTING TEXT COMPRESSION DRIVING PART AND METHOD FOR COMPRESSING INPUT TEXT COLUMN 失效
公开(公告)号：JPH06208453A
公开(公告)日：1994-07-26
申请号：JP19458393
申请日：1993-08-05
申请人： XEROX CORP
发明人： RONARUDO EMU KAPURAN , JIYON TEI MATSUKUSUUERU ZA SAA
IPC分类号： G06F5/00 , G06F17/21 , H03M7/42 , G06F15/20
摘要： PURPOSE: To provide a text compressing technique by which the compression ratio can be improved. CONSTITUTION: An uncompressed text train is resolved into tokens and the next token is extracted (step 50). The extracted token is transmitted to a WNM (word-number mapper) set and a word-number pair peculiar to the token is generated (step 52) and the token is encoded based on the word-number pair (step 55). In general, the number of used nibles is decided from the number of WNMs for indicating order. Encoded tokens indicating one nible to many nibles in which WNMs exist in sets are outputted to a compression database for insertion (step 57). Then an un-coded database is inspected for the presence of any other token (step 60). When a token exists in the database, the process returns to step 50 in which the next token is extracted. The above-mentioned steps are repeated until all tokens in an input train are encoded.

5. 发明专利

JPH05242142A 失效
公开(公告)号：JPH05242142A
公开(公告)日：1993-09-21
申请号：JP30272692
申请日：1992-11-12
申请人： XEROX CORP
发明人： EMU MAAGARETSUTO UIZUGOTSUTO , SUTEIIBUN SHII BAGUREI , DAN ESU BURUUMUBAAGU , PAA KURISUCHIYAN HARUBORUSEN , DANIERU PII HIYUTSUTENROTSUHIY , TOTSUDO EI KASU , RONARUDO EMU KAPURAN , RAMAANA BII RAO
IPC分类号： G06K9/20 , G06F17/28 , G06F17/30 , G06K9/72 , G06F15/38 , G06F15/40
摘要： PURPOSE: To provide a method and a device for abridging and summarizing an undecoded document picture without converting it into an optical character code. CONSTITUTION: The document picture is made into pieces concerning respective pages in a stage 20 and the position of the picture unit of the stage and its space relevancy are inspected in the stage 25. The picture unit which is not discriminated in the stage 30 is evaluated without decoding the picture unit under classification or without referring to decoded picture data based on the evaluation of a prescribed morphological picture characteristic in the stage 40. Evaluation is executed in the discrimination stage 41 of the morphological picture characteristic and the characteristics are compared in the stage 42 concerning the respective picture units as against the morphological picture characteristic which is obtained or selected by another picture unit. In the stage 50, the identified important picture unit is processed and the secondary and auxiliary document picture where the contents of a supply source document picture are reflected is generated.

6. 发明专利

JPH07129367A W/N AND N/W MAPPING METHOD 失效
公开(公告)号：JPH07129367A
公开(公告)日：1995-05-19
申请号：JP26742893
申请日：1993-10-26
申请人： XEROX CORP
发明人： RONARUDO EMU KAPURAN , MAATEIN KEI
IPC分类号： G06F5/00
摘要： PURPOSE: To improve the mapping efficiency by containing branch skip information showing the number of the ending of suffix inside a 2nd set to be used for mapping concerning a word, which is not finished by one piece of suffix ending in a 1st set, and providing the number of ending of the word in a 1st branch data item by accessing data items. CONSTITUTION: A 1st transition I reaches a big branch, but a 2nd transition R reaches the other big branch. Therefore, in the case of comparing the word with word list expressions or passing through all the branches from the 1st transition from a state 12 for counting the number of ending of suffix in the branch, it takes a long time. For example, a state 16 has the 1st transition P to reach the big branch but the 2nd transition S consists of an extremely small branch. The information required for skipping the branch and the number of ending of suffix in the branch are contained in the word list data so that the branch skip is enabled. Thus, the number of ending of suffix in the branch is not counted but increased/decreased to the count of execution.

7. 发明专利

JPH076214A METHOD FOR CHANGE OF CHARACTER AT INSIDE OD IMAGE 失效
公开(公告)号：JPH076214A
公开(公告)日：1995-01-10
申请号：JP40587990
申请日：1990-12-25
申请人： XEROX CORP
发明人： SUTEIIBUN SHII BAGURII , RONARUDO EMU KAPURAN , UEIRANDO AARU HITSUKUSU , DANIERU DEIBUISU
IPC分类号： G06K9/03 , G06F17/21 , G06K9/00 , G06K9/72 , G06T11/60 , G06F17/27
摘要： PURPOSE: To change the character of a text appearing in an image by providing an intermediate two-dimensional arrangement composed of words and providing a two-dimensional arrangement for defining the character from this intermediate arrangement. CONSTITUTION: The image area of a frame 10 is scanned and the two-dimensional arrangement of data, with which the respective items of data are binary values or any other digital values showing respective positions in the image area, is provided. When it is recognized that a word 12 is not correct, a correct spelling is displayed on an image output device like a frame 20 and a user selects a word 22 and the spelling of 'weird' while setting a pointer 24 on the wrong word. At such a time point, data defining characters more than one for the word 22 are provided from the two-dimensional arrangement defining the image area. Further, the device provides data for defining the word selected in the frame 20 while using data found out in frames 30, 40 and 50 and provides data for defining the corrected edition of the text, in which the word 12 is replaced with a word 72 while using data defining a word 62 and the 1st two-dimensional arrangement.

8. 发明专利

JPH05250408A 失效
公开(公告)号：JPH05250408A
公开(公告)日：1993-09-28
申请号：JP30272792
申请日：1992-11-12
申请人： XEROX CORP
发明人： EMU MAAGARETSUTO UIZUGOTSUTO , UIRIAMU NIYUUMAN , SUTEIIBUN SHII BAGUREI , DANIERU PII HIYUTSUTENROTSUHIY , RONARUDO EMU KAPURAN , TOTSUDO EI KASU , PAA KURISUCHIYAN HARUBORUSEN , JIYON SHIIRII BURAUN , MAATEIN KEI
IPC分类号： G06F17/24 , G06F17/27 , G06F17/28 , G06F17/30 , G06K9/20 , G06F15/38 , G06F15/40
摘要： PURPOSE: To provide a method for electronic document processing, to improve a device and to provide a device for retrieving complementary data corresponding to the important part of a selected document without decoding any document. CONSTITUTION: The 1st state of image processing technique is document image analysis, the document image is divided 20 into non-decoded information, and the positional/spatial relation of image units on a certain page is investigated 25. Next, a discriminating stage 30 is executed and the image unit having information contents not enough to be effective for evaluating the contents of theme is identified. For this evaluation, discrimination 41 of image form characteristics and comparison 42 for image form characteristics are required. A reader lets a user read and select the document or any fractacle of this document and obtain the interactive retrieval of the document and complementary data.

9. 发明专利

JPH02157968A UNIFICATION OF LOGIC OR 失效
公开(公告)号：JPH02157968A
公开(公告)日：1990-06-18
申请号：JP26793389
申请日：1989-10-13
申请人： XEROX CORP
发明人： JIYONTEII MATSUKUSUUERU ZA SAA , RONARUDO EMU KAPURAN
IPC分类号： G06F9/44 , G06F17/27 , G06F17/28
摘要： PURPOSE: To efficiently resolve an OR system by unifying a first data unit with a second data unit and accessing a context identifier related to the first data unit and using the context identifier to decide whether first and second data units can be unified or not. CONSTITUTION: With respect to unification of two data structures 22 and 24, the data unit of one data structure is unified with that of the other to obtain one new data unit where two unified data units are coupled. In the case of OR, context identifiers C1 to C4 which discriminate the combinations of OR branches are used. Each OR branch in combinations can be discriminated by a peculiar identifier. An OR branch and OR can be discriminated by, for example, peculiar numbers, and these two numbers are combined to form a choice, and this choice is combined with another to from the context identifiers C1 to C4 . Thus, the OR system is efficiently resolved.

10. 发明专利

JPH09153047A TOKEN DIVISION METHOD FOR TEXT 失效
公开(公告)号：JPH09153047A
公开(公告)日：1997-06-10
申请号：JP21546296
申请日：1996-07-26
申请人： XEROX CORP
发明人： RONARUDO EMU KAPURAN
IPC分类号： G06F17/22 , G06F17/27
摘要： PROBLEM TO BE SOLVED: To effectively divide a text in limited storage space by generating the output for every pinch point, guaranteeing that the output in each pinch point is a proper output and deleting the storage of the proceedings up to the time by an output created word. SOLUTION: The text recognized by an OCR 300 is delivered to a general purpose computer 90 and a token division is performed for the text. At first, the present new unprocessed configuration is generated by a configuration storage unit 40. This concficuration becomes the present unprocessed configuration, the extended state becomes the value of the next extended state variable and the output node becomes an empty node. The present unprocessed configuration is selected and is processed. The configuration storage unit 40 is checked and this check is repeated till the present unprocessed configuration is elimianted. When the present configuration is eliminated, all the present configurations are released.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式