CN1025764C - Characters recognition method and system - Google Patents
Characters recognition method and system Download PDFInfo
- Publication number
- CN1025764C CN1025764C CN 92103651 CN92103651A CN1025764C CN 1025764 C CN1025764 C CN 1025764C CN 92103651 CN92103651 CN 92103651 CN 92103651 A CN92103651 A CN 92103651A CN 1025764 C CN1025764 C CN 1025764C
- Authority
- CN
- China
- Prior art keywords
- character
- stroke
- stroke unit
- unit
- annexation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 18
- 239000000835 fiber Substances 0.000 claims description 17
- 230000008878 coupling Effects 0.000 claims description 16
- 238000010168 coupling process Methods 0.000 claims description 16
- 238000005859 coupling reaction Methods 0.000 claims description 16
- 239000000284 extract Substances 0.000 claims description 10
- 230000002093 peripheral effect Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 5
- 230000015572 biosynthetic process Effects 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 238000012163 sequencing technique Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 4
- 239000000463 material Substances 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 4
- 230000002349 favourable effect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 10
- 241000233805 Phoenix Species 0.000 description 2
- 238000013475 authorization Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 244000062793 Sorghum vulgare Species 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- IXSZQYVWNJNRAL-UHFFFAOYSA-N etoxazole Chemical compound CCOC1=CC(C(C)(C)C)=CC=C1C1N=C(C=2C(=CC=CC=2F)F)OC1 IXSZQYVWNJNRAL-UHFFFAOYSA-N 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 235000019713 millet Nutrition 0.000 description 1
- 238000007639 printing Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Images
Landscapes
- Character Discrimination (AREA)
Abstract
The present invention relates to a character recognition method and a system. Stroke characteristics of a character picture are extracted, and characters are classified, matched and identified by directly making use of the stroke characteristics. Structure meaning of the characters is expressed by adopting a framework form, important influence strokes and stroke connections in a framework are emphasized, strokes with less action is omitted, and a necessary comparison condition for stroke direction of variable permission and similar character discrimination are presented. The present invention is favorable for protruding distinction between the characters, and processes of matching and identification are simplified. Compared with the present universal character identification technology, the present invention has high identification rate and adaptability.
Description
The present invention relates to a kind of character identifying method and system, be particularly useful for discerning the recognition methods of handwritten Chinese character and multi-font printing Chinese character.
Some character recognition systems of having developed both at home and abroad mainly adopt picture dot to character image to distribute to extract characteristic parameter, and serve as according to classifying and mate the character identifying method of discerning with this parameter.For example, the character recognition system of Chinese patent authorization on February 8th, 1989 bulletin CN1003257B, the disclosed technology of Chinese patent authorization bulletin CN1010512B on November 21 nineteen ninety.
Therefore, common technology has following problem:
1. can not directly reflect the architectural feature of character, thereby ignore the essential characteristic that the stroke structure constitutes as character.
2. be difficult to the discrimination that reaches high under the situation of large character set.
3. it is very difficult to distinguish plesiomorphism or the baroque character of stroke.
4. under the handwritten character situation, font is write and is altered a great deal, and the characteristic parameter that is extracted is dispersed big, and needs to adopt the high dimensional feature vector.
The objective of the invention is to create a kind of character identifying method, make every effort to extract exactly the stroke feature of character image, fully reflect the structure essence of character; The stroke structure meaning of a word that directly utilizes character is to character classification and coupling identification; Use the structure meaning of a word of knowledge representation character, reach the coupling identifying of abbreviated character, improve the accuracy of the similar character of identification and the adaptive faculty of recognition methods.
Character identifying method involved in the present invention comprises: it is first step that the page scan of writing character is obtained character image; Character image binaryzation, character cutting and specification turn to second step; The stroke architectural feature that extracts the character binaryzation dot matrix is a third step; Try to achieve the characteristic of division sign indicating number by architectural feature and under determining, be categorized as the 4th step; The character model of architectural feature and affiliated classification mated and identification be the 5th step; Recognition result transferred to as seen be output as the 6th step.
Described third step comprises:
1. the charcter topology pattern can be decomposed into metacharacter, stroke and stroke unit three spermotypes as pattern integral body.Metacharacter is the character of structure character.Stroke is decomposed into straight-line segment and is stroke unit.Stroke unit is the lowermost level subpattern, and as the structural motif of describing character pattern, its architectural feature comprises stroke unit centre coordinate, length, direction and annexation.
2. character pattern is done once simply scanning, detect each picture dot on 8 directions with the situation that is connected of adjacent pixel, it is divided into top, terminal, bonding pad or common stroke element and the corresponding symbol of mark of stroke, thereby character pattern plane (CDP) converted to character picture dot attribute plane (CAP).
3. except that the picture dot that belongs to the bonding pad, on CAP, be in the picture dot of marginal point, calculate its " | ", "-", "/", " continuous picture dot number en on \ " four direction, the direction of en maximum is got the fiber principal direction of making this marginal point.En value on principal direction is called fibre length, and the picture dot that connects on the fibre length is composed with the corresponding weights of principal direction.The fiber of each marginal point may intersect the formation interwoven region, and its direction weights of the picture dot of interwoven region add up.All marginal points can be tried to achieve character fibrous structure chart (CFP) after finishing aforementioned calculation.
4. contrast the direction character of CAP bonding pad, remove the noise fiber among the CFP, to belong to " | ", "-", "/", " fiber of \ " four direction places V, h, four planes of s, b respectively, can try to achieve centre coordinate, length and the direction of each stroke unit.
5. utilize end points and the bonding pad feature of CAP, can calculate the annexation of stroke unit in conjunction with stroke unit centre coordinate, length and the direction of having asked.
Described the 4th step comprises:
1. four corner characteristics of application character peripheral structure and four limit features are carried out the description and the classification of peripheral structure as the characteristic of division of character on two levels.Four corner characteristics and four limit features by known character are set up the dictionary of presorting.
2. (CSP) is the center with four angles on plane on the stroke plane of character, four jiaos of nearest stroke units of detection range.
3. judge the stroke unit direction attribute of nearest angle point, and be divided into six types of horizontal, vertical, left-falling strokes, right-falling stroke, angle, friendship, compose, be called the angle sign indicating number with respective coding.The sign indicating number string of being made up of four angle sign indicating numbers constitutes first characteristic of division of character.
4. on CSP, draw ray by the center, by scanning clockwise, the polygon that acquisition ray and character outermost layer stroke unit are formed extracts the salient point that it surpasses a certain threshold value as the peripheral profile of character, and the salient point number of counting each limit is respectively tried to achieve second characteristic of division of the sign indicating number string formation character on four limits.
5. search in the dictionary of presorting and wait become literate Fu Si angle sign indicating number and the identical similar character code of four-sides code, finish the 4th step.
Described the 5th step:
1. the charcter topology meaning of a word adopts frame form's knowledge representation, expresses each character pattern by character frame.In framework, the whole stroke unit that constitutes character is packet sequencing on h, v, four planes of s, b respectively, and lists the discrimination condition of stroke unit feature between necessary stroke annexation and the similar character.Each the stroke unit that participates in packet sequencing in character frame is by stroke unit frame description.Stroke unit framework is expressed normal direction, center and the length of stroke unit.In addition, give the weight of this stroke and the distortion direction of permission.Weight in necessary annexation in the character frame and the stroke unit framework belongs to utilization knowledge representation, emphasize to recognition result the stroke unit of material impact and annexation thereof are arranged and ignore those redundancies or influence little composition.The distortion direction of similar character discrimination condition and permission makes identifying can take in the huge character set of complex structure and quantity the nuance of stroke structure between the identification kinds of characters into account, can have good adaptive faculty to the font that changes all the time again.
2. take out the similar character model of presorting, carry out search matched, computation attribute distance with the stroke unit feature that accords with of waiting to become literate successively,, otherwise think that it fails to match if distance thinks that less than a certain threshold value the match is successful.So process is carried out until end on four sub-planes of stroke unit of each model successively.
3. according to the Weighted distance of the weight calculation stroke meta-attribute of stroke unit framework appointment.The stroke unit that charcter topology is played a crucial role is owing to the nuance that has the highest weight to be convenient to distinguish the intercharacter stroke, and influencing little stroke unit has less weight, thereby reaches the purpose of ignoring redundant stroke.
In the stroke unit that do not become of coupling if there is the sub-plane of the sample tolerable distortion direction, that turn to respective direction search matched.
5. the annexation to necessity detects, and does not satisfy and withdraws from the matching candidate row when this requires.
6. detect stroke unit and compare and similar character discrimination condition, withdraw from the matching candidate row when not meeting the demands.
7. distance ordering from small to large pressed in the total distance of coupling all characters in threshold range, takes out minimum several conduct identification candidate, do not handle to refuse knowledge if there is the identification candidate.
The distinct advantages that the present invention has can be summarized as follows:
Thereby accurately extract the essential characteristic that the stroke architectural feature has fully reflected character.Directly utilize stroke feature to describe the structural framework of character and adapt to the many variations of character form, realize character classification and coupling identification with the stroke property vector.To the structure meaning of a word model use frame form's of character knowledge representation, both be convenient to emphasize important stroke or stroke annexation, can ignore again the little stroke of identification character influence, very help outstanding intercharacter difference and simplify the coupling identifying.Express the discrimination condition of similar character in the framework, made the trickle stroke difference of identification intercharacter become possibility, for example: wind, phoenix; Scholar, soil; Billows, calumniate ... thereby, greatly improved the discrimination of character.In the stroke framework, give the direction that allows distortion, make the dirigibility and the adaptive faculty of identification significantly improve.Compare with existing technology, both avoided in the statistical method because of adopting high dimensional feature to exist the difficulty of feature selecting and pattern separability aspect to limit the raising of discrimination.Also avoided structural approach to be difficult to adapt to the changeable defective of character form.
Embodiments of the invention are made up of image-text scanner, micro-mainframe computer, display, printer, magnetic tape station and relevant interface board.Scanner comprises that the hand-hold scanning various types is all applicable.Micro-mainframe computer uses dos operating system the most general.The dispensable equipment of magnetic tape station can be used as the expansion or the reserve of mainframe memory and freely selects for use.The principle of work of system progressively illustrates in conjunction with following accompanying drawing.
Description of drawings:
Fig. 1 is the square construction drawing of the embodiment of the invention
Fig. 2 is that architectural feature extracts workflow diagram
Fig. 3 is the example that architectural feature extracts
Fig. 4 is that stroke unit annexation is described
Fig. 5 is the workflow diagram of presorting
Fig. 6 is four corner characteristics code tables
Fig. 7 is a character frame
Fig. 8 is the condition ordering structure figure of stroke unit
Fig. 9 is a stroke unit condition ordering workflow diagram
Figure 10 is a stroke unit framework
Figure 11 is the coupling identification workflow diagram of utilization knowledge elicitation
Figure 12 is a sub-plane h stroke unit coupling workflow diagram.
Fig. 1 is the system block diagrams of embodiment, and the character of writing on paper scans the page with image-text scanner, and every page of scanning obtains an images file, converts binaryzation (0,1) dot matrix to by selected gray threshold, in interface board deposits computing machine in.Initial row by page segmentation program module search dot matrix, the row sum, prefix and number of words are finished the cutting of word automatically, after handling, normalization obtains the dot matrix (for example 32 * 32 or 64 * 64 character patterns) of each character, extract the stroke feature of each character pattern, classify, mate so discern the character pattern of this character to the machine of being stored in all identification finish, represent recognition result with internal code.Show at last or print the recognition result of writing alphabet on specimen page, perhaps proceed necessary editor with the standard font.
Fig. 2 is the process flow diagram that architectural feature extracts, with the starting point of the character pattern (CDP) after the normalization processing as this flow process, the row and column of scanning CDP, detection is that 1 continuous picture dot is counted X in two direction values of row and column, write down the maximum X of occurrence number as stroke width wi, be expert at and column direction when measuring the not enough wi of continuous image element, use this picture dot of " | " and "-" mark respectively with stroke width.Whether be 0, be 0 to belong to left end point as the left side if detecting it in the both sides of "-" picture dot, be labeled as " W ".As the right side is 0 to belong to right endpoint, is labeled as " E ".Upper and lower two sides at " | " picture dot detect whether it is 0, and the top is 0 to belong to upper extreme point and be labeled as " N ", and the below is 0 to belong to lower extreme point and be labeled as " S ".All use small letter English character mark neither "-" also is not the picture dot of " | " in proper order by its regional coordinate in CDP.The zone of this English character mark is the bonding pad of stroke, and calculates the feature of this bonding pad.Each picture dot of CDP is called character picture dot attribute plane (CAP) by above-mentioned requirements by the different attributes such as top, terminal, bonding pad or common picture dot of promptly giving stroke after the designated symbols mark, and Fig. 3 illustrates the architectural feature extraction example that written character " adjoins " word.Wherein the upper left side is CAP figure, and the below is the bonding pad mark sheet, and first row are that sequence number, secondary series are that row coordinate, the 5th, six row that bonding pad code name, third and fourth row are respectively the initial sum terminations are respectively the horizontal ordinates of initial sum termination.Last row are connection features of bonding pad, and connection features is shown in Fig. 4 with coded representation.Each marginal point to CAP, except that the picture dot of bonding pad, on row, column, oblique, the right tiltedly four direction in a left side, calculate its picture dot number of non-0 continuously, get its picture dot and count the fiber principal direction of the direction of maximum as this marginal point, the picture dot number that connects on the principal direction is a fibre length, and each picture dot is composed with the corresponding weights of principal direction.The fiber of each marginal point may intersect the formation interwoven region, and its direction weights of interwoven region picture dot add up.All marginal points are promptly tried to achieve character fibrous structure chart (CFP) after finishing aforementioned calculation.Remove the noise fiber of interwoven region, to belong to row, column, a left side tiltedly, the right tiltedly fiber of four direction places h, v, four planes of s, b can try to achieve centre coordinate, length and the direction of each stroke unit respectively, utilize end points and the connection features of CAP to try to achieve the annexation of stroke unit again, thereby obtain the entire infrastructure feature of character.The top right plot of Fig. 3 shows the example of " adjoining " word architectural feature.
Fig. 5 is the workflow diagram of presorting.On the stroke plane of character, be the center with four angles on plane, four jiaos of nearest stroke units of detection range.Judge the direction attribute of this stroke unit, they are divided into horizontal, vertical, left-falling stroke, right-falling stroke, angle, friendship and empty seven types.Their coding is called the angle sign indicating number as shown in Figure 6, and the sign indicating number string of being made up of four angle sign indicating numbers constitutes first characteristic of division of character.On the stroke plane, draw ray by the center again, by scanning clockwise, the polygon that acquisition ray and character outermost layer stroke unit are formed is as the peripheral profile of character, extract the salient point that it surpasses a certain threshold value, the salient point number that calculates each limit respectively is as the limit sign indicating number, and four limit sign indicating numbers constitute second characteristic of division that the four-sides code string is character.Search the dictionary of presorting by four-sides code and four jiaos of sign indicating numbers, obtain similar character code.
Fig. 7 is the framework of expressing charcter topology meaning of a word model, the wherein ε of subscripting
iRepresent i stroke unit, packet sequencing on h, v, four sub-faces of s, b respectively, Fig. 8 are the condition ordering structure figure of stroke unit, and sort criteria can be with reference to Fig. 9.Necessary annexation Ω
MnBe meant the annexation that must satisfy between m stroke of this character unit and n the stroke unit, for example: " husband " word must be the relation that intersects between the first horizontal pen and the perpendicular pen, day does not then have this requirement.Stroke unit compares notch then in order to distinguish the difference of inner stroke length comparison of character or direction, for example: soil, scholar; My god, die young, similar character discrimination condition judges then that a certain stroke unit lacks or when existing, the shift direction of candidate characters, for example: wind, phoenix; Beam, fine strain of millet or the like.Figure 10 is that stroke unit framework is expressed each ε of rule unit among Fig. 7
iArchitectural feature.Comprise stroke unit normal direction be quantified as horizontal, vertical, cast aside, press down for four direction respectively with h, v, s, b representative; Centre coordinate (the x of stroke unit
o, y
o)
iWith stroke length.Give the direction ε that this stroke allows distortion in the framework
' iWith structure ratio wi, the former makes matching process flexibly and the adaptive faculty that the raising system changes font, and the latter then gives top priority to what is the most important and simplifies coupling.Fig. 7 and Figure 10 form the structural model of system.Figure 11 illustrates the coupling identification workflow diagram of utilization knowledge elicitation.Figure 12 is certain sub-plane stroke unit coupling workflow diagram.Take out corresponding character model one by one according to the given similar character code of presorting from knowledge base, the condition sequencer program module of being represented by Fig. 9 sorts to the stroke unit of trying to achieve.Mate successively between organizing after the stroke in model unit and the stroke unit that treats character learning symbol are earlier in the group successively on h, v, four sub-planes of s, b, the attributive distance between the stroke thinks that the match is successful during less than defined threshold δ, otherwise thinks that it fails to match.If it fails to match, whether searching character stroke unit mates downwards, as do not have coupling may, get next model stroke unit and mate.This process is performed until last model stroke unit.If all the match is successful or stroke unit coupling finishes in model stroke unit, then according to the attributive distance of the whole strokes of weight calculation of appointment, distance is thought within threshold value △ scope the time can list matching candidate in.If existence allows the distortion direction, turn to the sub-plane of the sample search matched of respective direction in the stroke unit that coupling does not become in the model, method is identical.Further detect the annexation the Ω whether symbol of waiting to become literate satisfies appointment for the character model of listing matching candidate in
Mn, for example: husband, sky; Power, cutter; , husband and Li overlapping relation all are necessary, do not satisfy and withdraw from the matching candidate row when this requires.If exist in the model framework stroke unit relatively requirement check whether meet the demands, that does not satisfy comparison condition withdraws from candidate's row.Repeating above-mentioned matching ratio finishes until all classification Model Matching.Mating is first candidate by distance order arrangement from small to large as what discern candidate arrangement first place apart from all characters in threshold range always, generally is taken as recognition result.If not having the identification candidate then handles to refuse to know.
Claims (1)
1, a kind of character identifying method, it is first step that the page scan of writing character is obtained character image; Character image binaryzation, character cutting and specification turn to second step: the stroke architectural feature that extracts the character binaryzation dot matrix is a third step; Try to achieve the characteristic of division sign indicating number by architectural feature and under determining, be categorized as the 4th step; The character model of architectural feature and affiliated classification mated and identification be the 5th step; Recognition result transferred to as seen be output as the 6th step, feature of the present invention is:
Described third step comprises:
(1) the charcter topology pattern can be decomposed into metacharacter, stroke and stroke unit three spermotypes as pattern integral body.Metacharacter is the character of structure character, and stroke is decomposed into straight-line segment and is stroke unit.Stroke unit is the lowermost level subpattern, and as the structural motif of describing character pattern, its architectural feature comprises stroke unit centre coordinate, length, direction and annexation.
(2) character pattern is done once simply scanning, detect each picture dot on 8 directions with the situation that is connected of adjacent pixel, it is divided into top, terminal, bonding pad or common stroke element and the corresponding symbol of mark of stroke, thereby character pattern plane (CDP) converted to character picture dot attribute plane (CAP).
(3) except that the picture dot that belongs to the bonding pad, on CAP, be in the picture dot of marginal point, calculate its " | ", " one ", "/", " " continuous picture dot number en on the four direction, the direction of en maximum is got the fiber principal direction of making this marginal point.En value on principal direction is called fibre length, and the picture dot that connects on the fibre length is composed with the corresponding weights of principal direction.The fiber of each marginal point may intersect the formation interwoven region, and its direction weights of the picture dot of interwoven region add up.All marginal points can be tried to achieve character fibrous structure chart (CFP) after finishing aforementioned calculation.
(4) direction character of contrast CAP bonding pad, remove the noise fiber among the CFP, to belong to " | ", " one ", "/", " " fiber of four direction places V, h, four planes of s, b respectively, can try to achieve centre coordinate, length and the direction of each stroke unit.
(5) utilize end points and the bonding pad feature of CAP, can calculate the annexation of stroke unit in conjunction with stroke unit centre coordinate, length and the direction of having asked.
Described the 4th step comprises:
(1) four corner characteristics of application character peripheral structure and four limit features are carried out the description and the classification of peripheral structure as the characteristic of division of character on two levels.Four corner characteristics and four limit features by known character are set up the dictionary of presorting.
(2) (CSP) is the center with four angles on plane on the stroke plane of character, four jiaos of nearest stroke units of detection range.
(3) judge the stroke unit direction attribute of nearest angle point, and be divided into six types of horizontal, vertical, left-falling strokes, right-falling stroke, angle, friendship, compose, be called the angle sign indicating number with respective coding.The sign indicating number string of being made up of four angle sign indicating numbers constitutes first characteristic of division of character.
(4) on CSP, draw ray by the center, by scanning clockwise, the polygon that acquisition ray and character outermost layer stroke unit are formed is as the peripheral profile of character, extract the salient point that it surpasses a certain threshold value, the salient point number of counting each limit is respectively tried to achieve second characteristic of division of the sign indicating number string formation character on four limits.
(5) search in the dictionary of presorting and wait become literate Fu Si angle sign indicating number and the identical similar character code of four-sides code, finish the 4th step.
Described the 5th step:
(1) the charcter topology meaning of a word adopts frame form's knowledge representation, expresses each character pattern by character frame.In framework, the whole stroke unit that constitutes character is packet sequencing on h, v, four planes of s, b respectively, and lists the discrimination condition of stroke unit feature between necessary stroke annexation and the similar character.Each the stroke unit that participates in packet sequencing in character frame is by stroke unit frame description.Stroke unit framework is expressed normal direction, center and the length of stroke unit.In addition, give the weight of this stroke and the distortion direction of permission.Weight in necessary annexation in the character frame and the stroke unit framework belongs to utilization knowledge representation, emphasize to recognition result the stroke unit of material impact and annexation thereof are arranged and ignore those redundancies or influence little composition.The distortion direction of similar character discrimination condition and permission makes identifying can take in the huge character set of complex structure and quantity the nuance of stroke structure between the identification kinds of characters into account, can have good adaptive faculty to the font that changes all the time again.
(2) take out the similar character model of presorting, carry out search matched, computation attribute distance with the stroke unit feature that accords with of waiting to become literate successively,, otherwise think that it fails to match if distance thinks that less than a certain threshold value the match is successful.So process is carried out until end on four sub-planes of stroke unit of each model successively.
(3) according to the Weighted distance of the weight calculation stroke meta-attribute of stroke unit framework appointment.The stroke unit that charcter topology is played a crucial role is owing to the nuance that has the highest weight to be convenient to distinguish the intercharacter stroke, and influencing little stroke unit has less weight, thereby reaches the purpose of ignoring redundant stroke.
(4) in the stroke unit that do not become of coupling if there is the sub-plane of the sample tolerable distortion direction, that turn to respective direction search matched.
(5) annexation to necessity detects, and does not satisfy and withdraws from the matching candidate row when this requires.
(6) detect stroke unit and compare and similar character discrimination condition, withdraw from the matching candidate row when not meeting the demands.
(7) distance ordering from small to large pressed in the total distance of coupling all characters in threshold range, takes out minimum several conduct identification candidate, do not handle to refuse knowledge if there is the identification candidate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 92103651 CN1025764C (en) | 1992-05-12 | 1992-05-12 | Characters recognition method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 92103651 CN1025764C (en) | 1992-05-12 | 1992-05-12 | Characters recognition method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1066335A CN1066335A (en) | 1992-11-18 |
| CN1025764C true CN1025764C (en) | 1994-08-24 |
Family
ID=4940353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 92103651 Expired - Fee Related CN1025764C (en) | 1992-05-12 | 1992-05-12 | Characters recognition method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1025764C (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1036685C (en) * | 1995-06-20 | 1997-12-10 | 贾东升 | Information recording and reproducing apparatus |
| CN102254204A (en) * | 2011-06-03 | 2011-11-23 | 吴林 | Coding and decoding method for graphemic code |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1317664C (en) * | 2004-01-17 | 2007-05-23 | 中国科学院计算技术研究所 | Confused stroke order library establishing method and on-line hand-writing Chinese character identifying and evaluating system |
| JP5071914B2 (en) * | 2005-02-28 | 2012-11-14 | ザイ デクマ アクチボラゲット | Recognition graph |
| CN1332348C (en) * | 2005-09-23 | 2007-08-15 | 清华大学 | Blocks letter Arabic character set text dividing method |
| CN101436248B (en) * | 2007-11-14 | 2012-10-24 | 佳能株式会社 | Method and equipment for generating text character string according to image |
| CN101436254B (en) * | 2007-11-14 | 2013-07-24 | 佳能株式会社 | Image processing method and image processing equipment |
| CN102024138B (en) * | 2009-09-15 | 2013-01-23 | 富士通株式会社 | Character identification method and character identification device |
| CN102096662A (en) * | 2010-12-06 | 2011-06-15 | 无敌科技(西安)有限公司 | Code conversion method |
| CN103366716B (en) * | 2012-03-31 | 2016-03-30 | 华为终端有限公司 | The compression of character and dot matrix word library and decompress(ion) method and apparatus in dot matrix word library |
| CN103870516B (en) * | 2012-12-18 | 2019-10-25 | 北京三星通信技术研究有限公司 | Retrieve the method for image, paint in real time reminding method and its device |
| CN107633250B (en) * | 2017-09-11 | 2023-04-18 | 畅捷通信息技术股份有限公司 | Character recognition error correction method, error correction system and computer device |
| CN112487985A (en) * | 2020-11-30 | 2021-03-12 | 江苏云控软件技术有限公司 | Method for positioning water gauge of ship |
-
1992
- 1992-05-12 CN CN 92103651 patent/CN1025764C/en not_active Expired - Fee Related
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1036685C (en) * | 1995-06-20 | 1997-12-10 | 贾东升 | Information recording and reproducing apparatus |
| CN102254204A (en) * | 2011-06-03 | 2011-11-23 | 吴林 | Coding and decoding method for graphemic code |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1066335A (en) | 1992-11-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1025764C (en) | Characters recognition method and system | |
| US6335986B1 (en) | Pattern recognizing apparatus and method | |
| CN1122243C (en) | Automatic language identification system for multilingual optical character recognition | |
| Kahan et al. | On the recognition of printed characters of any font and size | |
| US5995659A (en) | Method of searching and extracting text information from drawings | |
| CN107194400A (en) | A kind of finance reimbursement unanimous vote is according to picture recognition processing method | |
| CN1163841C (en) | On-line hand writing Chinese character distinguishing device | |
| US20020154815A1 (en) | Character recognition device and a method therefore | |
| US5048113A (en) | Character recognition post-processing method | |
| RU2640322C2 (en) | Methods and systems of effective automatic recognition of symbols | |
| CN106778717A (en) | A kind of test and appraisal table recognition methods based on image recognition and k nearest neighbor | |
| US5926564A (en) | Character recognition method and apparatus based on 0-1 pattern representation of histogram of character image | |
| CN100485711C (en) | Computer identification and automatic inputting method for hand writing character font | |
| CN115880566A (en) | Intelligent marking system based on visual analysis | |
| KR19980086524A (en) | Pattern extraction device | |
| Mozaffari et al. | ICDAR 2009 handwritten Farsi/Arabic character recognition competition | |
| Yin et al. | Handwritten text line extraction based on minimum spanning tree clustering | |
| CN110032999A (en) | A kind of low resolution licence plate recognition method that Hanzi structure is degenerated | |
| CN1641681A (en) | Method for rapid inputting character information for mobile terminal with pickup device | |
| Tofani et al. | Segmentation of text from color map images | |
| CN1052203A (en) | Off-line Handwritten Chinese Recognition system and recognition methods thereof | |
| JP3083609B2 (en) | Information processing apparatus and character recognition apparatus using the same | |
| KR910005390B1 (en) | Method for processing document automatically and recognizing english characters | |
| CN85100085A (en) | Recognition method of printed Chinese character recognition device | |
| Raza et al. | Recognition of facsimile documents using a database of robust features |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C06 | Publication | ||
| PB01 | Publication | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |