CN111401351A - Segmentation method based on vertical character positioning expansion - Google Patents
Segmentation method based on vertical character positioning expansion Download PDFInfo
- Publication number
- CN111401351A CN111401351A CN202010149621.1A CN202010149621A CN111401351A CN 111401351 A CN111401351 A CN 111401351A CN 202010149621 A CN202010149621 A CN 202010149621A CN 111401351 A CN111401351 A CN 111401351A
- Authority
- CN
- China
- Prior art keywords
- answer
- answers
- character
- positioning
- picture
- 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.)
- Granted
Links
- 230000011218 segmentation Effects 0.000 title claims abstract description 26
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000005520 cutting process Methods 0.000 claims abstract description 26
- 238000012360 testing method Methods 0.000 claims abstract description 10
- 238000013527 convolutional neural network Methods 0.000 claims description 6
- 238000003825 pressing Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 6
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 206010061619 Deformity Diseases 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
Images
Classifications
-
- 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
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- 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
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
- G06V30/28—Character recognition specially adapted to the type of the alphabet, e.g. Latin alphabet
- G06V30/287—Character recognition specially adapted to the type of the alphabet, e.g. Latin alphabet of Kanji, Hiragana or Katakana characters
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Character Input (AREA)
Abstract
The invention relates to a segmentation method based on vertical character positioning expansion, which comprises the following steps: (1) text line positioning: carrying out text line positioning on the test paper, wherein the coordinate obtained by positioning is the determined text line height; (2) and (3) cutting an answer area: cutting the answer area of the test paper according to the height of the text line determined in the step (1), and separating out a character picture containing the answer; if the answers in the character pictures containing the answers separated in the step (2) are adhered or the answers are not completely in the answer area, performing adhesion expansion or incomplete expansion on the text line, and then cutting to obtain the character pictures containing complete answers; (3) character recognition: carrying out batch identification on the character pictures containing the answers obtained by expanding in the step (2) to obtain identification answers; (4) and (5) judging a result: and comparing the recognition answers obtained by batch recognition with the answers in the answer library to obtain whether the recognition answers are correct answers or not.
Description
Technical Field
The invention relates to the technical field of text image processing, in particular to a segmentation method based on vertical character positioning expansion.
Background
Machine vision is an important branch of artificial intelligence, has been developed rapidly in recent years, has been widely applied in various industries, greatly improves human productivity and production modes, reduces labor of people, relieves workload of people, and brings great convenience and improvement to life styles of people.
In the field of education, computer vision technology has been applied to automated review systems in a comprehensive manner, which can help teachers perform automated review of test papers. The time for the teacher to manually correct the test paper is greatly shortened, and the working efficiency of the teacher is improved. Although the automatic reading and amending system is nearly perfect, the practical application still faces many problems, wherein the OCR technology largely and completely depends on the normalization degree of the practical writing, for example, when students answer on the scroll, the written characters are stuck and exceed the answering area due to the limited range of the answering area. This causes serious problems in positioning answers, even failure to accurately position the answers of students, and finally failure to correctly identify the answers.
In order to accurately position answers, the application provides a segmentation method based on vertical character positioning expansion, which is used for solving the problem of vertical adhesion or non-separable staggered characters, and performing multiple times of answer area expansion during segmentation so as to obtain a character picture containing complete answers and achieve the aim of accurate identification.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a segmentation method based on vertical character positioning expansion, which solves the problem of vertically adhered or inseparable staggered characters, and performs multiple times of answer area expansion during segmentation so as to obtain a character picture containing complete answers, improve the accuracy of recognition and realize the purpose of accurate recognition.
In order to solve the technical problems, the invention adopts the technical scheme that: the segmentation method based on vertical character positioning expansion specifically comprises the following steps:
(1) text line positioning: carrying out text line positioning on the test paper, wherein the coordinate obtained by positioning is the determined text line height;
(2) and (3) cutting an answer area: cutting an answer area of the test paper according to the height of the text line determined in the step (1), and separating a character picture containing an answer; if the answers in the character pictures containing the answers separated in the step (2) are adhered or the answers are not completely in the answer area, performing adhesion expansion or incomplete expansion on the text lines, and then cutting to obtain the character pictures containing the complete answers;
(3) character recognition: performing batch recognition on the character pictures containing the answers obtained by the expansion in the step (2) by adopting a Convolutional Neural Network (CNN) to obtain recognition answers;
(4) and (5) judging a result: and comparing the recognition answers obtained by batch recognition with the answers in the answer library to obtain whether the recognition answers are correct answers or not.
As a preferred technical solution of the present invention, if the answers in the separated character picture containing the answers in step (2) are adhered, the specific steps of performing adhesion expansion on the text line and then cutting to obtain the character picture containing the answers are as follows:
s21-1, firstly, determining the expansion height through the space between the text line and the upper text line and the lower text line;
s21-2, expanding upwards according to the height, and cutting out a new picture A;
s21-3, extending downwards according to the height, and cutting out a new picture B;
and S21-4 expands at least once both upwards and downwards at the same time according to the expanded height of the steps S21-1 and S21-2, cuts out a new picture C, and obtains the character picture containing the answer. Here it can be extended multiple times over an extended height range. The traditional connected domain can not be used for positioning accurate characters, and the cutting result directly according to the height of the text line can not be complete characters; therefore, when the adhesion occurs, the adhesion expansion and cutting are required.
As a preferred embodiment of the present invention, if the answer in the character picture including the answer separated in step (2) is not completely located in the character characterized by the upper and lower structure in the answer area, the character is located as two or more connected domains.
As a preferred technical solution of the present invention, if the answer in the separated character picture including the answer in step (2) is not completely in the answer area, the specific steps of performing a default expansion on the text line and then cutting the text line to obtain the character picture including the answer are as follows:
s22-1, performing up-and-down expansion on the answer area, performing connected domain positioning on the expanded picture to obtain a plurality of connected domains, and then obtaining the number of characters appearing in the current answer area according to the answer library to perform further judgment;
s22-1-1, if the number of the positioned connected domains is the same as the number of characters appearing in the current answer area, no further processing is carried out;
s22-1-2, if the number of the positioned connected domains is more than the number of characters appearing in the current answer area, and the characters with the upper and lower structural relations exist in the handwritten characters, fusing the contents of the positioned connected domains, and combining the answer area and the expanded contents into an identification area.
As a preferred technical solution of the present invention, in the step (4), if one of the plurality of identification answers matches the answer in the answer library, it is determined that the answer to the question is correct; if the recognition answers are not consistent with the answers in the answer library, judging the answer of the question as a wrong answer.
As a preferred technical solution of the present invention, in the step S22-1, connected domain positioning is performed on the extended picture, and the specific steps of the connected domain positioning are as follows: firstly, traversing a first point P (x, y) with a pixel value in a picture according to rows and columns, giving one label to the first point P, and then pressing all foreground pixels adjacent to the pixel point into a stack; secondly, popping up a stack top pixel, endowing the same label to the stack top pixel, then pressing all foreground pixels adjacent to the stack top pixel into a stack, and repeating the steps until the stack is empty, so that a certain communicated area in the picture is obtained; and finally, repeating the steps to complete the traversal of the whole picture, and finally obtaining the connected regions of all characters in the picture, thereby realizing the positioning of the text characters.
Compared with the prior art, the invention has the beneficial effects that: the segmentation method based on vertical character positioning expansion solves the problem of vertically adhered or inseparable staggered characters, and multiple times of answer area expansion are carried out during segmentation, so that a character picture containing complete answers is obtained, and the identification accuracy is improved.
Drawings
The technical scheme of the invention is further described by combining the accompanying drawings as follows:
FIG. 1 is a flow chart of a segmentation method based on vertical character position extension of the present invention;
FIG. 2 is a schematic diagram of the height of a text line in step (1) of the segmentation method based on vertical character location extension according to the present invention;
FIG. 3 is a diagram of the occurrence of blocking during clipping in step (2) of the segmentation method based on vertical character positioning extension according to the present invention;
FIG. 4 is a graph of the disfigurement phenomenon cropped out in step (2) of the segmentation method based on vertical character positioning expansion according to the present invention;
FIG. 5 is a schematic diagram of the upward expansion of step S21-2 in the segmentation method based on vertical character positioning expansion according to the present invention;
FIG. 6 is a schematic diagram of the upward expansion of step S21-3 in the segmentation method based on vertical character positioning expansion according to the present invention;
FIG. 7 is a schematic diagram of the upward expansion of step S21-4 in the segmentation method based on vertical character positioning expansion according to the present invention;
FIG. 8 is a schematic diagram of the step S22-1-2 expansion fusion in the segmentation method based on vertical character positioning expansion of the present invention.
Detailed Description
For the purpose of enhancing the understanding of the present invention, the present invention will be described in further detail with reference to the accompanying drawings and examples, which are provided for the purpose of illustration only and are not intended to limit the scope of the present invention.
Example (b): as shown in fig. 1, the segmentation method based on vertical character positioning expansion specifically includes the following steps:
(1) text line positioning: carrying out text line positioning on the test paper, wherein the coordinate obtained by positioning is the determined text line height; the heights of different text lines are different, and the heights of the text lines determine the height of image segmentation; as shown in fig. 2;
(2) and (3) cutting an answer area: cutting an answer area of the test paper according to the height of the text line determined in the step (1), and separating a character picture containing an answer; if the answers in the character pictures containing the answers separated in the step (2) are adhered or the answers are not completely in the answer area, performing adhesion expansion or incomplete expansion on the text lines, and then cutting to obtain the character pictures containing the complete answers; the separated answers may be stuck as shown in fig. 3, or the answers are not completely in the answer area as shown in fig. 4; the answers cut in this step do not necessarily meet the final recognition requirements; that is, a character picture not containing a complete answer may appear;
if the answers in the separated character pictures containing the answers are adhered in the step (2), the step of cutting the character pictures containing the answers after adhering and expanding the text line comprises the following specific steps:
s21-1, firstly, determining the expansion height through the space between the text line and the upper text line and the lower text line;
s21-2, expanding upwards according to the height, and cutting out a new picture A; as shown in figure 5 of the drawings,
s21-3, extending downwards according to the height, and cutting out a new picture B; as shown in figure 6 of the drawings,
s21-4, expanding at least once both upwards and downwards at the same time according to the expanded height of the steps S21-1 and S21-2, cutting out a new picture C, as shown in FIG. 7, and obtaining the character picture containing the answer; the method can be expanded for multiple times within the expanded height range, so that a character picture containing a complete answer is obtained;
if the answer in the character picture containing the answer is not completely positioned in the character with the characteristic of the answer area as the upper and lower structure in the step (2), positioning the character into two or more connected domains; characters are not only likely to be adhered, but also likely to be incompletely positioned, which generally occurs in Chinese characters, and is characterized in that characters with an upper structure and a lower structure are positioned into two or more connected domains, as shown in fig. 3, a black character is positioned into two parts, and if only one connected domain is taken, a recognition error is caused;
if the answer in the separated character picture containing the answer in the step (2) is not completely in the answer area, performing the incomplete extension on the text line and then cutting to obtain the character picture containing the answer, specifically comprising the following steps:
s22-1, performing up-and-down expansion on the answer area, performing connected domain positioning on the expanded picture to obtain a plurality of connected domains, and then obtaining the number of characters appearing in the current answer area according to the answer library to perform further judgment; in the step S22-1, connected domain positioning is performed on the expanded picture, and the connected domain positioning specifically includes the steps of: firstly, traversing a first point P (x, y) with a pixel value in a picture according to rows and columns, giving one label to the first point P, and then pressing all foreground pixels adjacent to the pixel point into a stack; secondly, popping up a stack top pixel, endowing the same label to the stack top pixel, then pressing all foreground pixels adjacent to the stack top pixel into a stack, and repeating the steps until the stack is empty, so that a certain communicated area in the picture is obtained; and finally, repeating the steps to complete the traversal of the whole picture, and finally obtaining the connected regions of all characters in the picture, thereby realizing the positioning of the text characters.
S22-1-1, if the number of the positioned connected domains is the same as the number of characters appearing in the current answer area, no further processing is carried out;
s22-1-2, if the number of the positioned connected domains is more than the number of characters appearing in the current answer area, and the characters with the upper and lower structural relations exist in the handwritten characters, fusing the contents of the positioned connected domains, and combining the answer area and the expanded contents into an identification area;
as shown in fig. 8, the lower four points of "black" are separated from the upper half, so that the complete character cannot be located, so that the upper part and the lower part of "black" are expanded according to the position of the answer area, and finally the upper half and the lower half of "black" are fused, that is, the answer area and the expanded content are merged into one identification area;
(3) character recognition: performing batch recognition on the character pictures containing the answers obtained by the expansion in the step (2) by adopting a Convolutional Neural Network (CNN) to obtain recognition answers;
(4) and (5) judging a result: comparing the recognition answers obtained by batch recognition with answers in the answer library to obtain whether the recognition answers are correct answers or not;
in the step (4), if one of the multiple recognition answers matches with the answer in the answer library, judging that the answer of the question is correct; if the recognition answers are not consistent with the answers in the answer library, judging the answer of the question as a wrong answer.
It is obvious to those skilled in the art that the present invention is not limited to the above embodiments, and it is within the scope of the present invention to adopt various insubstantial modifications of the method concept and technical scheme of the present invention, or to directly apply the concept and technical scheme of the present invention to other occasions without modification.
Claims (6)
1. A segmentation method based on vertical character positioning extension is characterized by comprising the following steps:
(1) text line positioning: carrying out text line positioning on the test paper, wherein the coordinate obtained by positioning is the determined text line height;
(2) and (3) cutting an answer area: cutting an answer area of the test paper according to the height of the text line determined in the step (1), and separating a character picture containing an answer; if the answers in the character pictures containing the answers separated in the step (2) are adhered or the answers are not completely in the answer area, performing adhesion expansion or incomplete expansion on the text lines, and then cutting to obtain the character pictures containing the complete answers;
(3) character recognition: performing batch recognition on the character pictures containing the answers obtained by the expansion in the step (2) by adopting a Convolutional Neural Network (CNN) to obtain recognition answers;
(4) and (5) judging a result: and comparing the recognition answers obtained by batch recognition with the answers in the answer library to obtain whether the recognition answers are correct answers or not.
2. The segmentation method based on vertical character positioning extension according to claim 1, wherein if the answers in the separated character pictures containing answers are sticky in step (2), the step of obtaining the character pictures containing answers by clipping after sticky extension of the text line comprises the specific steps of:
s21-1, firstly, determining the expansion height through the space between the text line and the upper text line and the lower text line;
s21-2, expanding upwards according to the height, and cutting out a new picture A;
s21-3, extending downwards according to the height, and cutting out a new picture B;
and S21-4 expands at least once both upwards and downwards at the same time according to the expanded height of the steps S21-1 and S21-2, cuts out a new picture C, and obtains the character picture containing the answer.
3. The segmentation method based on vertical character positioning extension as claimed in claim 1, wherein if the answer in the character picture including the answer separated in step (2) is not completely located in the character characterized by the upper and lower structure of the answer area, the character is positioned into two or more connected domains.
4. The segmentation method based on vertical character positioning extension as claimed in claim 3, wherein if the answer in the separated character picture containing the answer in step (2) is not completely in the answer area, the step of performing the incomplete extension on the text line and then cutting the text line to obtain the character picture containing the answer comprises the specific steps of:
s22-1, performing up-and-down expansion on the answer area, performing connected domain positioning on the expanded picture to obtain a plurality of connected domains, and then obtaining the number of characters appearing in the current answer area according to the answer library to perform further judgment;
s22-1-1, if the number of the positioned connected domains is the same as the number of characters appearing in the current answer area, no further processing is carried out;
s22-1-2, if the number of the positioned connected domains is more than the number of characters appearing in the current answer area, and the characters with the upper and lower structural relations exist in the handwritten characters, fusing the contents of the positioned connected domains, and combining the answer area and the expanded contents into an identification area.
5. The segmentation method based on vertical character positioning extension as claimed in claim 3, wherein in the step (4), if one of the recognition answers matches the answer comparison in the answer library, the answer of the question is determined to be correct; if the recognition answers are not consistent with the answers in the answer library, judging the answer of the question as a wrong answer.
6. The segmentation method based on vertical character positioning extension according to claim 4, wherein the step S22-1 is to perform connected component positioning on the extended picture, and the connected component positioning specifically comprises: firstly, traversing a first point P (x, y) with a pixel value in a picture according to rows and columns, giving one label to the first point P, and then pressing all foreground pixels adjacent to the pixel point into a stack; secondly, popping up a stack top pixel, endowing the stack top pixel label, then pressing all foreground pixels adjacent to the stack top pixel into a stack, and repeating the steps until the stack is empty, so that a certain communicated area in the picture is obtained; and finally, repeating the steps to complete the traversal of the whole picture, and finally obtaining the connected regions of all characters in the picture, thereby realizing the positioning of the text characters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010149621.1A CN111401351B (en) | 2020-03-06 | 2020-03-06 | Segmentation method based on vertical character positioning expansion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010149621.1A CN111401351B (en) | 2020-03-06 | 2020-03-06 | Segmentation method based on vertical character positioning expansion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111401351A true CN111401351A (en) | 2020-07-10 |
| CN111401351B CN111401351B (en) | 2023-07-11 |
Family
ID=71430551
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010149621.1A Active CN111401351B (en) | 2020-03-06 | 2020-03-06 | Segmentation method based on vertical character positioning expansion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111401351B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156865A (en) * | 2010-12-14 | 2011-08-17 | 上海合合信息科技发展有限公司 | Handwritten text line character segmentation method and identification method |
| CN106096564A (en) * | 2016-06-17 | 2016-11-09 | 福建网龙计算机网络信息技术有限公司 | A kind of mathematics corrects method automatically |
| CN109408803A (en) * | 2018-08-29 | 2019-03-01 | 蓝舰信息科技南京有限公司 | A method of it semantic understanding for subjective item natural language and corrects |
| CN110378310A (en) * | 2019-07-25 | 2019-10-25 | 南京红松信息技术有限公司 | A kind of automatic generation method of the handwriting samples collection based on answer library |
-
2020
- 2020-03-06 CN CN202010149621.1A patent/CN111401351B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156865A (en) * | 2010-12-14 | 2011-08-17 | 上海合合信息科技发展有限公司 | Handwritten text line character segmentation method and identification method |
| CN106096564A (en) * | 2016-06-17 | 2016-11-09 | 福建网龙计算机网络信息技术有限公司 | A kind of mathematics corrects method automatically |
| CN109408803A (en) * | 2018-08-29 | 2019-03-01 | 蓝舰信息科技南京有限公司 | A method of it semantic understanding for subjective item natural language and corrects |
| CN110378310A (en) * | 2019-07-25 | 2019-10-25 | 南京红松信息技术有限公司 | A kind of automatic generation method of the handwriting samples collection based on answer library |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111401351B (en) | 2023-07-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108171297B (en) | Answer sheet identification method | |
| CN108932508B (en) | Method and system for intelligently identifying and correcting subjects | |
| CN109840519B (en) | Self-adaptive intelligent bill identification and input device and application method thereof | |
| US20210020059A1 (en) | Answer evaluation method, answer evaluation system, electronic device, and medium | |
| CN106980857B (en) | Chinese calligraphy segmentation and recognition method based on copybook | |
| CN108509988B (en) | Test paper score automatic statistical method and device, electronic equipment and storage medium | |
| CN113343815B (en) | Method for identifying form in PDF (Portable document Format) and other format documents | |
| CN110705630A (en) | Semi-supervised learning type target detection neural network training method, device and application | |
| CN105260751A (en) | Character identification method and system | |
| CN112446262A (en) | Text analysis method, text analysis device, text analysis terminal and computer-readable storage medium | |
| WO2023045298A1 (en) | Method and apparatus for detecting table lines in image | |
| CN111881659B (en) | Processing method and system of table picture, readable storage medium and computer equipment | |
| CN110837793A (en) | Intelligent recognition handwriting mathematical formula reading and amending system | |
| CN111144270B (en) | Neural network-based handwritten text integrity evaluation method and evaluation device | |
| CN112116840B (en) | Job correction method and system based on image recognition and intelligent terminal | |
| CN107220610A (en) | A kind of subjective item fraction recognition methods applied to marking system | |
| CN114581932A (en) | Picture table line extraction model construction method and picture table extraction method | |
| CN111401351A (en) | Segmentation method based on vertical character positioning expansion | |
| JP4756447B2 (en) | Teaching material processing apparatus, teaching material processing method, and teaching material processing program | |
| US8385811B1 (en) | System and method for processing forms using color | |
| CN110084117B (en) | Document table line detection method and system based on binary image segmentation projection | |
| CN113051457A (en) | Image-text extraction method and terminal | |
| CN113486786B (en) | Automatic operation modifying system | |
| CN113065477B (en) | Intelligent classroom textbook content identification method | |
| CN112686253A (en) | Screen character extraction system and method for electronic whiteboard |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |