CN111797827A - Automatic OCR recognition method for character direction mixed arrangement - Google Patents
Automatic OCR recognition method for character direction mixed arrangement Download PDFInfo
- Publication number
- CN111797827A CN111797827A CN202010421724.9A CN202010421724A CN111797827A CN 111797827 A CN111797827 A CN 111797827A CN 202010421724 A CN202010421724 A CN 202010421724A CN 111797827 A CN111797827 A CN 111797827A
- Authority
- CN
- China
- Prior art keywords
- rotation
- original file
- character
- ocr recognition
- coordinates
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/242—Aligning, centring, orientation detection or correction of the image by image rotation, e.g. by 90 degrees
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Input (AREA)
Abstract
The invention belongs to the technical field of OCR recognition, and provides an automatic OCR recognition method for character direction mixed arrangement, which comprises the following steps: s101: identifying coordinates of each vertex of the original file; s102: recognizing a character area, initial coordinates and diagonal coordinates of the character area; s103: filling the identified character areas with colors; s104: performing OCR recognition on the rotated file, recording recognized characters, and recognizing a rotated character area, initial coordinates and diagonal coordinates of the rotated character area; s105: judging whether the rotation times N reach the maximum value (360/A); s106: and outputting the recognized characters, drawing a text region rectangular frame of the original file through each vertex coordinate of the original file, and restoring the positions of the characters in the original file. The automatic OCR recognition method for character direction mixed arrangement has the advantage that recognition omission caused by the limitation of an OCR model is greatly reduced.
Description
Technical Field
The invention relates to the technical field of OCR (optical character recognition), in particular to an automatic OCR (optical character recognition) method for character direction mixed arrangement.
Background
In recent years, with the rapid development of big data technology and artificial intelligence technology, OCR recognition technology has been advanced. Based on the big data sample and the artificial intelligence technology, the normal print character recognition rate of OCR recognition, especially OCR recognition of print materials, is very close to 100%.
However, recognition of OCR also presents some challenges in the real engineering application field or some specific scenarios. For example, when a mobile phone shoots or a scanner scans, the direction is not well controlled, so that the picture is rotated by 90 degrees as a whole, and special processing is required when characters are greatly inclined, for example, the inclined angle is 90 degrees, 180 degrees or 270 degrees, or otherwise, characters lying sideways or upside down are difficult to be correctly recognized through OCR.
The current common treatment methods include: the method comprises the following steps of adding character samples in all directions, training different sample models of 90 degrees, 180 degrees, 270 degrees and the like, and processing by using different models during recognition, wherein the method has the following problems:
1. a large number of samples need to be constructed in the early stage, so that the cost is high;
2. if different models are obtained by training aiming at different character directions, different models are required to be selected for recognizing characters when OCR recognition is carried out, and the problem of larger delay is caused in the process of selecting different models for recognition;
3. if characters in all directions are trained in one model, the model is extremely large, and the OCR recognition efficiency of the system is greatly reduced.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides the automatic OCR recognition method for the character direction mixed arrangement, which can greatly reduce the missing caused by the limitation of an OCR model aiming at the mixed arrangement original file.
In order to solve the technical problems, the invention provides the following technical scheme:
an automatic OCR recognition method for character direction mixed arrangement comprises the following steps:
s101: identifying each vertex coordinate of an original file, setting a rotation unit A of the original file, initializing the rotation times N of the original file to be 0, and initializing the rotation times N to be {1, 2., (360/A) };
s102: recognizing a character region and an initial coordinate (X) of the character region0，Y0) And diagonal coordinates (X)1，Y1) And recording the recognized characters;
s103: carrying out color filling on the identified character region, carrying out A-angle rotation on the original file, adding 1 to the rotation times N, and recording the accumulated rotation angle of the original file as A x N;
s104: OCR recognition is carried out on the rotated file, recognized characters are recorded, and a character area after rotation and initial coordinates (M) of the character area after rotation are recognized0N，N0N) And diagonal coordinate (M)1N，N1N)；
S105: judging whether the rotation times N reach the maximum value (360/A), if so, performing step S106; otherwise, repeating step S103;
s106: outputting recognized characters, drawing a text region rectangular frame of the original document through each vertex coordinate of the original document, and according to different rotation units A × N of the original document and initial coordinates (M) corresponding to the different rotation units A × N0N，N0N) And diagonal coordinate (M)1N，N1N) And restoring the position of the characters in the original file.
Further, in the step S106, the initial coordinates (M) corresponding to different rotation units a × N and different rotation units a × N of the original file are used0N，N0N) And diagonal coordinate (M)1N，N1N) The method for restoring the position of the characters in the original file comprises the following steps:
s10601: establishing initial coordinates (M) after rotation according to different rotation units A x N of original file0N，N0N) With initial coordinates (X)0，Y0) S functional relationship between1And S2，Wherein (M)0N，N0N)＝S1(X0) And (M)0N，N0N)＝S2(Y0) And diagonal coordinates after rotation (M)1N，N1N) And diagonal coordinate (X)1，Y1) S functional relationship between3And S4Wherein (M)1N，N1N)＝S3(X1) And (M)1N，N1N)＝S4(Y1)；
S10602: will function as S1And S2Inverse transformation is carried out to obtain X0And Y0Wherein X is0＝S1 -1(M0N，N0N) And Y0＝S2 -1(M0N，N0N) (ii) a Will function as S3And S4Inverse transformation is carried out to obtain X1And Y1Wherein X is1＝S3 -1(M1N，N1N) And Y1＝S4 -1(M1N，N1N)。
Further, the rotation unit A of the original file is 90 degrees, and when the rotation angle is 90 degrees, N is01＝S1(X0)，M01＝S2(Y0)，N11＝S3(X1)，M11＝S4(Y1) Wherein S is1＝-1，S2＝1，S3＝-1，S41 is ═ 1; when the rotation angle is 180 degrees, M02＝S1(X0)，N02＝S2(Y0)，M12＝S3(X1)，N12＝S4(Y1) Wherein S is1＝1，S2＝-1，S3＝1，S4-1; when the rotation angle is 270 degrees, N03＝S1(X0)，M03＝S2(Y0)，N13＝S3(X1)，M13＝S4(Y1) Wherein S is1＝1，S2＝-1，S3＝1，S4＝-1。
Further, the a-angle rotation in step S103 includes clockwise rotation and counterclockwise rotation.
Further, the rotated document, recognized character and character area initial coordinates (M)0N，N0N) And diagonal coordinate (M)1N，N1N) Are stored in a disk file system or memory.
According to the technical scheme, the invention has the beneficial effects that: through lasting A angle rotation to original file, simultaneously, original file is once rotatory, carry out once discernment to original file through OCR, discern original file forward's characters after rotatory in proper order, 360 degrees rotations are accomplished to original file, OCR can discern original file in the slope, reverse characters, need not newly-increased characters sample, need not newly-increased OCR recognition model, need not to train newly-increased characters sample and OCR recognition model, the original file that significantly reduces the mixed row when carrying out OCR recognition, because the restriction of model leads to the missing of characters.
Drawings
In order to more clearly illustrate the detailed description of the invention or the technical solutions in the prior art, the drawings that are needed in the detailed description of the invention or the prior art will be briefly described below. Throughout the drawings, like elements or portions are generally identified by like reference numerals. In the drawings, elements or portions are not necessarily drawn to scale.
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a schematic diagram of an original document according to the present invention;
FIG. 3 is a schematic view of an original document rotated 90 degrees according to the present invention;
FIG. 4 is a schematic view of an original document rotated 180 degrees according to the present invention;
FIG. 5 is a schematic diagram of an original document rotated 270 degrees according to the present invention;
FIG. 6 is a flowchart illustrating the step S106 of restoring the position of the text in the original file according to the present invention.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and therefore are only examples, and the protection scope of the present invention is not limited thereby.
Referring to fig. 1 to 5, the present embodiment provides an automatic OCR recognition method for character direction mixing, including the following steps:
s101: identifying each vertex coordinate of an original file, setting a rotation unit A of the original file, initializing the rotation times N of the original file to be 0, and the rotation times N to be {1, 2., (360/A) }, wherein the original file is a picture file or a PDF file, if the original file is the PDF file, the content of the PDF file needs to be processed page by page, and the rotation of the original file is the prior art, so the repeated description is omitted;
s102: recognizing a character region and an initial coordinate (X) of the character region0，Y0) And diagonal coordinates (X)1，Y1) And recording the recognized characters;
s103: filling colors in the identified character areas, rotating the original file by an angle A, adding 1 to the rotation times N, recording the accumulated rotation angle of the original file as A x N, wherein the filled colors can be black, white or non-character patterns;
s104: OCR recognition is carried out on the rotated file, recognized characters are recorded, and a character area after rotation and initial coordinates (M) of the character area after rotation are recognized0N，N0N) And diagonal coordinate (M)1N，N1N)；
S105: judging whether the rotation times N reach the maximum value (360/A), if so, performing step S106; otherwise, repeating step S103;
s106: outputting recognized characters, drawing a text region rectangular frame of the original document through each vertex coordinate of the original document, and according to different rotation units A × N of the original document and initial coordinates (M) corresponding to the different rotation units A × N0N，N0N) And diagonal coordinate (M)1N，N1N) And restoring the position of the characters in the original file.
In the practical use, the original file is continuously rotated by the angle A, meanwhile, the original file is rotated once, the original file is identified once through the OCR, forward characters after the original file is rotated are sequentially identified, 360-degree rotation is completed until the original file, the OCR can identify inclined and reverse characters in the original file, a character sample does not need to be newly added, an OCR identification model does not need to be newly added, training on the newly added character sample and the OCR identification model is not needed, and the omission of characters caused by limitation of the model when the original file which is mixed and arranged is subjected to OCR identification is greatly reduced.
Referring to fig. 6, in the step S106, the initial coordinates (M) corresponding to different rotation units a × N and different rotation units a × N of the original file0N，N0N) And opposite angleCoordinate (M)1N，N1N) The method for restoring the position of the characters in the original file comprises the following steps:
s10601: establishing initial coordinates (M) after rotation according to different rotation units A x N of original file0N，N0N) With initial coordinates (X)0，Y0) S functional relationship between1And S2，Wherein (M)0N，N0N)＝S1(X0) And (M)0N，N0N)＝S2(Y0) And diagonal coordinates after rotation (M)1N，N1N) And diagonal coordinate (X)1，Y1) S functional relationship between3And S4Wherein (M)1N，N1N)＝S3(X1) And (M)1N，N1N)＝S4(Y1)；
S10602: will function as S1And S2Inverse transformation is carried out to obtain X0And Y0Wherein X is0＝S1 -1(M0N，N0N) And Y0＝S2 -1(M0N，N0N) (ii) a Will function as S3And S4Inverse transformation is carried out to obtain X1And Y1Wherein X is1＝S3 -1(M1N，N1N) And Y1＝S4 -1(M1N，N1N)。
In practical use, the initial coordinate (X) before each rotation is recorded0，Y0) And diagonal coordinates (X)1，Y1) And initial coordinates after rotation (M)0N，N0N) And diagonal coordinate (M)1N，N1N) The functional relation is established between the characters, so that the characters which are inclined and reversed can be restored conveniently after the original file is identified, the identified characters are not lost and misplaced, and the method has fidelity.
In this embodiment, the rotation unit a of the original file is 90 degrees, and when the rotation angle is 90 degrees, N is01＝S1(X0)，M01＝S2(Y0)，N11＝S3(X1)，M11＝S4(Y1) Wherein S is1＝-1，S2＝1，S3＝-1，S41 is ═ 1; when the rotation angle is 180 degrees, M02＝S1(X0)，N02＝S2(Y0)，M12＝S3(X1)，N12＝S4(Y1) Wherein S is1＝1，S2＝-1，S3＝1，S4-1; when the rotation angle is 270 degrees, N03＝S1(X0)，M03＝S2(Y0)，N13＝S3(X1)，M13＝S4(Y1) Wherein S is1＝1，S2＝-1，S3＝1，S4＝-1。
In actual use, the original file is identified after being rotated once, and the rotation angle is 90 degrees at the moment, through the function S1，S2，S3And S4Carrying out inverse transformation to obtain the original coordinates of the recognized characters; the original file is identified after rotating twice, and the rotating angle is 180 degrees at the moment, through the function S1，S2，S3And S4Carrying out inverse transformation to obtain the original coordinates of the recognized characters; the original file is recognized after being rotated three times, and the rotation angle at this time is 270 degrees by using the function S1，S2，S3And S4And performing inverse transformation to obtain the original coordinates of the recognized characters. In addition, when the rotation unit A can be 30 degrees or 45 degrees, the original file is sequentially rotated through the rotation angle of 30 degrees or 45 degrees to perform character recognition, and the functional relation S is in the process1，S2，S3And S4And will vary accordingly.
In this embodiment, the angle a rotation in step S103 includes a clockwise rotation and a counterclockwise rotation, and the picture can be rotated from both clockwise and counterclockwise directions.
In this embodiment, the rotated document, the recognized text and the initial coordinates of the text area: (M0N，N0N) And diagonal coordinate (M)1N，N1N) Are stored in a disk file system or memory.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the present invention, and they should be construed as being included in the following claims and description.
Claims (5)
1. An automatic OCR recognition method for character direction mixed arrangement is characterized by comprising the following steps:
s101: identifying each vertex coordinate of an original file, setting a rotation unit A of the original file, initializing the rotation times N of the original file to be 0, and initializing the rotation times N to be {1, 2., (360/A) };
s102: recognizing a character region and an initial coordinate (X) of the character region0，Y0) And diagonal coordinates (X)1，Y1) And recording the recognized characters;
s103: carrying out color filling on the identified character region, carrying out A-angle rotation on the original file, adding 1 to the rotation times N, and recording the accumulated rotation angle of the original file as A x N;
s104: OCR recognition is carried out on the rotated file, recognized characters are recorded, and a character area after rotation and initial coordinates (M) of the character area after rotation are recognized0N，N0N) And diagonal coordinate (M)1N，N1N)；
S105: judging whether the rotation times N reach the maximum value (360/A), if so, performing step S106; otherwise, repeating step S103;
s106: outputting the recognized characters, and drawing a text region rectangle of the original document through each vertex coordinate of the original documentAnd framing based on different rotation units A × N of the original file and the initial coordinates (M) corresponding to the different rotation units A × N0N，N0N) And diagonal coordinate (M)1N，N1N) And restoring the position of the characters in the original file.
2. An automated OCR recognition method according to claim 1, wherein in step S106, the original document is subjected to different rotation units a × N and initial coordinates (M) corresponding to the different rotation units a × N0N，N0N) And diagonal coordinate (M)1N，N1N) The method for restoring the position of the characters in the original file comprises the following steps:
s10601: establishing initial coordinates (M) after rotation according to different rotation units A x N of original file0N，N0N) With initial coordinates (X)0，Y0) S functional relationship between1And S2，Wherein (M)0N，N0N)＝S1(X0) And (M)0N，N0N)＝S2(Y0) And diagonal coordinates after rotation (M)1N，N1N) And diagonal coordinate (X)1，Y1) S functional relationship between3And S4Wherein (M)1N，N1N)＝S3(X1) And (M)1N，N1N)＝S4(Y1)；
S10602: will function as S1And S2Inverse transformation is carried out to obtain X0And Y0Wherein X is0＝S1 -1(M0N，N0N) And Y0＝S2 -1(M0N，N0N) (ii) a Will function as S3And S4Inverse transformation is carried out to obtain X1And Y1Wherein X is1＝S3 -1(M1N，N1N) And Y1＝S4 -1(M1N，N1N)。
3. The writing instrument as claimed in claim 2The automatic OCR recognition method of direction mixed arrangement is characterized in that the rotation unit A of the original file is 90 degrees, and when the rotation angle is 90 degrees, N is01＝S1(X0)，M01＝S2(Y0)，N11＝S3(X1)，M11＝S4(Y1) Wherein S is1＝-1，S2＝1，S3＝-1，S41 is ═ 1; when the rotation angle is 180 degrees, M02＝S1(X0)，N02＝S2(Y0)，M12＝S3(X1)，N12＝S4(Y1) Wherein S is1＝1，S2＝-1，S3＝1，S4-1; when the rotation angle is 270 degrees, N03＝S1(X0)，M03＝S2(Y0)，N13＝S3(X1)，M13＝S4(Y1) Wherein S is1＝1，S2＝-1，S3＝1，S4＝-1。
4. An automated OCR recognition method of character direction mixing according to claim 1, wherein the a-degree rotation in step S103 includes clockwise rotation and counterclockwise rotation.
5. An automated OCR recognition method of character direction mixing according to claim 1 and characterized in that the rotated document, recognized characters and character areas are rotated initial coordinates (M)0N，N0N) And diagonal coordinate (M)1N，N1N) Are stored in a disk file system or memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010421724.9A CN111797827A (en) | 2020-05-18 | 2020-05-18 | Automatic OCR recognition method for character direction mixed arrangement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010421724.9A CN111797827A (en) | 2020-05-18 | 2020-05-18 | Automatic OCR recognition method for character direction mixed arrangement |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111797827A true CN111797827A (en) | 2020-10-20 |
Family
ID=72805912
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010421724.9A Pending CN111797827A (en) | 2020-05-18 | 2020-05-18 | Automatic OCR recognition method for character direction mixed arrangement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111797827A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112329777A (en) * | 2021-01-06 | 2021-02-05 | 平安科技（深圳）有限公司 | Character recognition method, device, equipment and medium based on direction detection |
| CN112418238A (en) * | 2020-12-09 | 2021-02-26 | 安徽吉秒科技有限公司 | Image character recognition method and device |
| CN113033380A (en) * | 2021-03-23 | 2021-06-25 | 建信览智科技(北京)有限公司 | Text labeling method |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09311904A (en) * | 1996-05-21 | 1997-12-02 | Sharp Corp | Character recognition device |
| JPH1021376A (en) * | 1996-06-27 | 1998-01-23 | Ricoh Co Ltd | Electronic filing device |
| JP2010072826A (en) * | 2008-09-17 | 2010-04-02 | Ricoh Co Ltd | Image processing apparatus, image processing method, program, and computer readable storage medium |
| CN108960229A (en) * | 2018-04-23 | 2018-12-07 | 中国科学院信息工程研究所 | One kind is towards multidirectional character detecting method and device |
| CN109636815A (en) * | 2018-12-19 | 2019-04-16 | 东北大学 | A kind of metal plate and belt Product labelling information identifying method based on computer vision |
| CN109685059A (en) * | 2018-11-06 | 2019-04-26 | 平安科技（深圳）有限公司 | Character image mask method, device and computer readable storage medium |
| CN109977723A (en) * | 2017-12-22 | 2019-07-05 | 苏宁云商集团股份有限公司 | Big bill picture character recognition methods |
| CN110490198A (en) * | 2019-08-12 | 2019-11-22 | 上海眼控科技股份有限公司 | Text orientation bearing calibration, device, computer equipment and storage medium |
| CN110516554A (en) * | 2019-07-31 | 2019-11-29 | 杭州电子科技大学 | A kind of more scene multi-font Chinese text detection recognition methods |
-
2020
- 2020-05-18 CN CN202010421724.9A patent/CN111797827A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09311904A (en) * | 1996-05-21 | 1997-12-02 | Sharp Corp | Character recognition device |
| JPH1021376A (en) * | 1996-06-27 | 1998-01-23 | Ricoh Co Ltd | Electronic filing device |
| JP2010072826A (en) * | 2008-09-17 | 2010-04-02 | Ricoh Co Ltd | Image processing apparatus, image processing method, program, and computer readable storage medium |
| CN109977723A (en) * | 2017-12-22 | 2019-07-05 | 苏宁云商集团股份有限公司 | Big bill picture character recognition methods |
| CN108960229A (en) * | 2018-04-23 | 2018-12-07 | 中国科学院信息工程研究所 | One kind is towards multidirectional character detecting method and device |
| CN109685059A (en) * | 2018-11-06 | 2019-04-26 | 平安科技（深圳）有限公司 | Character image mask method, device and computer readable storage medium |
| CN109636815A (en) * | 2018-12-19 | 2019-04-16 | 东北大学 | A kind of metal plate and belt Product labelling information identifying method based on computer vision |
| CN110516554A (en) * | 2019-07-31 | 2019-11-29 | 杭州电子科技大学 | A kind of more scene multi-font Chinese text detection recognition methods |
| CN110490198A (en) * | 2019-08-12 | 2019-11-22 | 上海眼控科技股份有限公司 | Text orientation bearing calibration, device, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 秦旭光;徐爱功;: "图像几何校正在OCR文字识别中的应用", 辽宁工程技术大学学报, vol. 25, no. 2, pages 64 - 66 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112418238A (en) * | 2020-12-09 | 2021-02-26 | 安徽吉秒科技有限公司 | Image character recognition method and device |
| CN112329777A (en) * | 2021-01-06 | 2021-02-05 | 平安科技（深圳）有限公司 | Character recognition method, device, equipment and medium based on direction detection |
| CN113033380A (en) * | 2021-03-23 | 2021-06-25 | 建信览智科技(北京)有限公司 | Text labeling method |
| CN113033380B (en) * | 2021-03-23 | 2024-03-26 | 金科览智科技(北京)有限公司 | Text labeling method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111797827A (en) | Automatic OCR recognition method for character direction mixed arrangement | |
| CN105989317B (en) | Two-dimensional code identification method and device | |
| CN109344820B (en) | Digital ammeter reading identification method based on computer vision and deep learning | |
| CN110647882A (en) | Image correction method, device, equipment and storage medium | |
| JPH04233676A (en) | Convolution filtering method for decoding self-locking symbol-shape code | |
| CN113592735A (en) | Text page image restoration method and system, electronic equipment and computer readable medium | |
| CN105447508A (en) | Identification method and system for character image verification codes | |
| US20010024520A1 (en) | Method and apparatus for table recognition, apparatus for character recognition, and computer product | |
| JP5541679B2 (en) | Image processing apparatus and method, and program | |
| CN111353956A (en) | Image restoration method and device, computer equipment and storage medium | |
| CN116009749A (en) | Handwritten character erasing method and system based on attention mechanism | |
| JP2004272798A (en) | Image reading device | |
| CN110414517B (en) | Rapid high-precision identity card text recognition algorithm used for being matched with photographing scene | |
| CN114463770A (en) | Intelligent question-cutting method for general test paper questions | |
| JP3989733B2 (en) | Correction of distortion in form processing | |
| CN110610163B (en) | Table extraction method and system based on ellipse fitting in natural scene | |
| WO2021088333A1 (en) | Image storage method, image reading method, image memory and storage medium | |
| JP3698867B2 (en) | Circular pattern determination method, apparatus and recording medium | |
| CN112565549A (en) | Book image scanning method | |
| CN112651926A (en) | Method and device for detecting cracks based on recursive attention mechanism | |
| CN113177556A (en) | Text image enhancement model, training method, enhancement method and electronic equipment | |
| CN111899184A (en) | Image defect repairing and neural network training method, device, equipment and system | |
| JP4576358B2 (en) | Character image processing apparatus and character image processing method | |
| JP2004110434A (en) | Image processing method and image processing device | |
| CN117423111B (en) | Paper manuscript extraction and correction method and system based on computer vision and deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |