CN115841670B - Operation wrong question collecting system based on image recognition - Google Patents
Operation wrong question collecting system based on image recognition Download PDFInfo
- Publication number
- CN115841670B CN115841670B CN202310102859.2A CN202310102859A CN115841670B CN 115841670 B CN115841670 B CN 115841670B CN 202310102859 A CN202310102859 A CN 202310102859A CN 115841670 B CN115841670 B CN 115841670B
- Authority
- CN
- China
- Prior art keywords
- outline
- text
- comparison
- unit
- data
- 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.)
- Active
Links
Images
Classifications
-
- 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
- Image Analysis (AREA)
Abstract
The invention relates to the field of image data processing, in particular to an operation error question collecting system based on image recognition, which is provided with a database module, a data acquisition module, a data processing module and a data correction module, acquires picture information uploaded by a user side, recognizes text contours corresponding to answers in the picture based on the picture information, compares the text contours with contour patterns stored in the database, determines characters corresponding to the text contours, considers associated characters of adjacent text contours when the coincidence ratio of the text contours to the contour patterns is lower, and re-acquires the coincidence ratio comparison result of the text contours and the contour patterns corresponding to the associated characters after adjusting a coincidence ratio comparison threshold value, thereby improving the recognition rate of the characters corresponding to the text contours.
Description
Technical Field
The invention relates to the field of image data processing, in particular to a job error question collecting system based on image recognition.
Background
With the continuous progress of image recognition technology, the related technology is applied to various fields, especially in the fields of offices and education, such as file scanning, file recognition, job correction and the like;
chinese patent publication No.: CN111242045a discloses a method and system for indicating error of automatic operation problem, the method comprises: collecting images to be judged of a to-be-corrected exercise book or test paper in the color reference wire frame; comparing the image to be judged with the standard problem set to find out a standard problem image with highest matching degree; each standard problem image contains standard answer characters, positions and sizes of each small problem answer block; obtaining a small question answer sub-image of each small question answer block in the image to be judged, identifying answer characters from the small question answer sub-image through a standard function, comparing the answer characters with standard answer characters of the corresponding small question answer blocks in the standard exercise image, and determining that the answer of each small question answer block in the image to be judged is wrong; and projecting and outputting a color question number or a wrong number at a corresponding position on a corresponding exercise book or test paper according to the question answer block of each small question in the image to be judged. The invention can realize automatic and high-speed correct and wrong answer indication.
There are problems in the prior art that,
in the prior art, under the condition that outline patterns of characters are not clear or are not easy to identify, word group formation of unrecognized characters is not considered according to adjacent character analysis, so that the identification rate of the characters is improved.
Disclosure of Invention
In order to solve the problem of low text recognition accuracy in the prior art, the invention provides an operation wrong question collecting system based on image recognition, which comprises the following steps:
the database module comprises a first storage unit for storing the association relation among the characters and a second storage unit for storing a plurality of character outlines, and the corresponding relation is pre-established between the character outlines and the characters;
the data acquisition module is connected with the user terminal and used for receiving the picture information sent by the user terminal;
the data processing module comprises a first data comparing unit, a second data comparing unit, a first analyzing unit and a second analyzing unit,
the first data comparison unit is connected with the data acquisition module and is used for receiving the picture information, extracting text outlines of answers in the picture information and screening the text outlines of the answers in the picture information based on definition parameters of the text outlines of the answers in the picture information;
the second data comparison unit is respectively connected with the first data comparison unit and the database module and is used for calculating the coincidence degree of the character outline screened by the first data comparison unit and each outline pattern stored in the database module, and comparing each coincidence degree with a coincidence degree comparison threshold value or a coincidence degree correction comparison threshold value so as to obtain a coincidence degree comparison result of the character outline screened by the first data comparison unit and each outline pattern;
the first analysis unit is respectively connected with the second data comparison unit and the database module, and is used for determining characters corresponding to the character outline screened by the first data comparison unit based on the coincidence degree sequencing of the character outline screened by the first data comparison unit and each outline pattern under a first coincidence degree comparison result, and generating text information according to all the determined characters;
the second analysis unit is respectively connected with the second data comparison unit and the database module and is used for determining characters corresponding to adjacent character outlines of the character outlines screened by the first data comparison unit under a second coincidence ratio comparison result, determining outline patterns corresponding to associated characters of the characters, and sending the outline patterns to the second data comparison unit so that the second data comparison unit can acquire a coincidence ratio comparison result of the character outlines screened by the first data comparison unit and the received outline patterns again after correcting the coincidence ratio comparison threshold;
and the checking module is connected with the first analysis unit and is used for comparing the text information generated by the first analysis unit with preset comparison answer information and judging whether the text information is wrong or not.
Further, the first contact ratio comparison result is that the contact ratio of the character outline screened by the first data comparison unit and at least one outline pattern is more than or equal to the contact ratio comparison threshold value or the contact ratio correction comparison threshold value;
and the second coincidence degree comparison result is that the coincidence degrees of the character outlines and all outline patterns screened by the first data comparison unit are smaller than the coincidence degree comparison threshold value or the coincidence degree correction comparison threshold value.
Further, the first data comparison unit calculates a definition parameter D corresponding to each text outline of the answer in the picture information according to formula (1),
in the formula (1), S represents the area of the text outline of the answer in the picture information, S0 represents the average area value of each text outline of the answer in the picture information, C represents the chromaticity value of the text outline of the answer in the picture information, and C0 represents the average chromaticity value of each text outline of the answer in the picture information.
Further, the first data comparison unit compares the text definition parameter D corresponding to the text outline of the answer in the picture information with a preset definition comparison parameter D1, and screens out the text outline of the answer in the picture information according to the comparison result,
if the comparison result meets a first preset condition, the first data comparison unit judges that the text outline of the answer in the picture information is screened out;
the first preset condition is that D is more than or equal to D1.
Further, the first analysis unit sorts the overlapping degree of the text outline screened by the first data comparison unit and each outline pattern under the first overlapping degree comparison result, and in the sorting result, the text corresponding to the outline pattern with the highest overlapping degree is used as the text corresponding to the text outline screened by the first data comparison unit.
Further, the second analyzing unit obtains the coincidence degree judging result of the adjacent character outline of the character outline screened by the first data comparing unit,
if any adjacent character outline accords with a first coincidence degree comparison result, the second analysis unit acquires characters corresponding to the adjacent character outline judged by the first analysis unit, and determines associated characters with the acquired characters based on the association relation among the characters stored in the first storage unit so as to record and generate an associated character set;
and if the adjacent character outlines all accord with the second coincidence degree comparison result, the second analysis unit judges that the characters corresponding to the character outlines screened by the first data comparison unit cannot be identified.
Further, the first storage unit constructs an association relationship between each text, wherein,
the first storage unit stores a plurality of words, determines words constituting each word, and establishes association relation between each word constituting the word for any word.
Further, the second data comparison unit calculates a discrete parameter E according to formula (2),
in the formula (2), G (i) represents an average value of the overlapping ratio of the outline pattern corresponding to the i-th associated text in the associated text set and the outline pattern corresponding to the remaining text, n represents the number of text in the associated text set, and n is an integer greater than zero.
Further, the second data comparing unit receives the related text set and determines outline patterns corresponding to the related text in the related text set, when the second data comparing unit obtains the coincidence degree comparison result of the text outline screened by the first data comparing unit and the outline patterns, the discrete parameter E is compared with the discrete parameter comparison parameter E0, the coincidence degree comparison threshold H0 is corrected according to the comparison result,
the first correction mode is to correct the overlap ratio comparison threshold H0 to an overlap ratio correction comparison threshold H according to a first preset correction parameter H1, and set H=H20-H1;
the second correction mode is to correct the overlap ratio comparison threshold H0 to an overlap ratio correction comparison threshold H according to a second preset correction parameter H2, and set H=H2;
wherein h1 is smaller than h2, E is smaller than E0 in the first correction mode, and E is larger than or equal to E0 in the second correction mode.
Further, the preset comparison answer information is a text pre-stored in a checking module, the checking module compares the text information with the preset comparison answer information, wherein,
if the text information is the same as the preset comparison answer information, the comparison module judges that the text information is correct,
and if the text information is different from the preset comparison answer information, the comparison module judges that the text information is wrong.
Compared with the prior art, the method and the device have the advantages that the picture information uploaded by the user side is obtained through the database module, the data acquisition module, the data processing module and the data correction module, the text outline corresponding to the answer in the picture is identified based on the picture information, the text outline is compared with the outline pattern stored in the database, the characters corresponding to the text outline are determined, when the coincidence degree of the text outline and the outline pattern is low, the associated characters of the adjacent text outline are considered, after the coincidence degree comparison threshold is adjusted, the coincidence degree comparison result of the text outline and the outline pattern corresponding to the associated characters is obtained again, and therefore the recognition rate of the characters corresponding to the text outline is improved.
In particular, the data processing module screens the character outline based on the definition parameters of the character outline, only judges the corresponding character of the screened character outline, in the practical situation, the character outline in the picture information uploaded by the user side has larger difference due to the difference of writing, and the character outline cannot be identified due to poor definition of part of the character outline or combination of the character outline with other character outlines.
In particular, according to the invention, different analysis units are called for data processing based on the coincidence degree comparison results of the character outlines and different outline patterns, so that the data processing pressure is shared, and when the coincidence degree of the character outlines and at least one outline pattern is larger than or equal to the coincidence degree comparison threshold value or the coincidence degree correction comparison threshold value, the character outlines at the moment have higher coincidence degree with the outline patterns stored in the database module, so that characters corresponding to the character outlines can be judged directly based on the sequencing of the coincidence degree, and the recognition efficiency of the character outlines is improved.
In particular, when the coincidence degree of the outline of the character and all outline patterns is smaller than the coincidence degree comparison threshold or the coincidence degree correction comparison threshold, the coincidence degree comparison threshold of the outline of the character and all outline patterns is not high, and the associated characters of the characters corresponding to the outline of the adjacent character need to be determined at the moment.
In particular, the correction amount for correcting the coincidence degree comparison threshold value is determined based on the corresponding discrete parameter of the associated text set, the coincidence degree between outline patterns corresponding to all associated text in the associated text set is higher under the condition that the discrete parameter is lower, and if the coincidence degree between any outline pattern and text outline is lower, the coincidence degree between other outline patterns and text outline is also possibly lower, so that the coincidence degree comparison threshold value needs to be reduced more, the second coincidence degree comparison result still appears after the coincidence degree comparison threshold value is prevented from being reduced, and the recognition rate of the text corresponding to the text outline is improved under the premise of improving the reliability through the process.
Drawings
FIG. 1 is a schematic diagram of a task fault collection system based on image recognition according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a data processing module according to an embodiment of the invention;
fig. 3 is a schematic diagram of a database module according to an embodiment of the invention.
Detailed Description
In order that the objects and advantages of the invention will become more apparent, the invention will be further described with reference to the following examples; it should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are merely for explaining the technical principles of the present invention, and are not intended to limit the scope of the present invention.
It should be noted that, in the description of the present invention, terms such as "upper," "lower," "left," "right," "inner," "outer," and the like indicate directions or positional relationships based on the directions or positional relationships shown in the drawings, which are merely for convenience of description, and do not indicate or imply that the apparatus or elements must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention.
Furthermore, it should be noted that, in the description of the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention can be understood by those skilled in the art according to the specific circumstances.
Referring to fig. 1, fig. 2, and fig. 3, which are schematic diagrams of an image recognition-based task error collection system, a data processing module and a database module according to an embodiment of the invention, the image recognition-based task error collection system of the invention includes:
the database module comprises a first storage unit for storing the association relation among the characters and a second storage unit for storing a plurality of character outlines, and the corresponding relation is pre-established between the character outlines and the characters;
the data acquisition module is connected with the user terminal and used for receiving the picture information sent by the user terminal;
the data processing module comprises a first data comparing unit, a second data comparing unit, a first analyzing unit and a second analyzing unit,
the first data comparison unit is connected with the data acquisition module and is used for receiving the picture information, extracting text outlines of answers in the picture information and screening the text outlines of the answers in the picture information based on definition parameters of the text outlines of the answers in the picture information;
the second data comparison unit is respectively connected with the first data comparison unit and the database module and is used for calculating the coincidence degree of the character outline screened by the first data comparison unit and each outline pattern stored in the database module, and comparing each coincidence degree with a coincidence degree comparison threshold value or a coincidence degree correction comparison threshold value so as to obtain a coincidence degree comparison result of the character outline screened by the first data comparison unit and each outline pattern;
the first analysis unit is respectively connected with the second data comparison unit and the database module, and is used for determining characters corresponding to the character outline screened by the first data comparison unit based on the coincidence degree sequencing of the character outline screened by the first data comparison unit and each outline pattern under a first coincidence degree comparison result, and generating text information according to all the determined characters;
the second analysis unit is respectively connected with the second data comparison unit and the database module and is used for determining characters corresponding to adjacent character outlines of the character outlines screened by the first data comparison unit under a second coincidence ratio comparison result, determining outline patterns corresponding to associated characters of the characters, and sending the outline patterns to the second data comparison unit so that the second data comparison unit can acquire a coincidence ratio comparison result of the character outlines screened by the first data comparison unit and the received outline patterns again after correcting the coincidence ratio comparison threshold;
and the checking module is connected with the first analysis unit and is used for comparing the text information generated by the first analysis unit with preset comparison answer information and judging whether the text information is wrong or not.
Specifically, the invention does not limit the specific structures of the database module, the data acquisition module, the data processing module and the data correction module, and the invention can be a functional program applied to a computer and can only meet the functions of data storage, data processing and data exchange.
Specifically, the invention is not limited to a specific form of obtaining the text outline of the answer in the picture information, and can determine the specific position of the answer according to the characteristics, for example, determine the answer position according to the underline position, further obtain the text outline of the answer position, or preset the specific position of the answer in the picture, further identify the text outline of the specific position, and replace according to specific needs by a person skilled in the art.
Specifically, the present invention is not limited to a specific way of calculating the coincidence ratio of the text outline and the outline pattern, and in the prior art, text examples are mostly represented in the spatial domain of the image by using any shape text detector, and the representation method based on the spatial domain can be divided into two types, namely, pixel mask representation and outline point sequence representation. Wherein, the pixel mask representation method may require a complex and time-consuming post-processing process, and the requirement for training sample size is often larger; the method has limited expression capability on the highly curved text, and because the Fourier coefficient representation can fit any closed curve in theory and the text outline is more concentrated on the low-frequency component, the problem can be well solved by representing the irregular scene text example in the Fourier domain, so the method can model the text example outline in the Fourier domain instead of the space domain by carrying out Fourier transformation on the text example outline by using Fourier transformation, and can stably and simply gradually approach any closed outline so as to calculate the coincidence degree; of course, the related calculating mode of the contour coincidence degree is the prior art, and other coincidence degree calculating modes can be selected by a person skilled in the art to calculate the coincidence degree of the text contour and the contour pattern in the invention.
Specifically, the first contact ratio comparison result is that the contact ratio of the character outline screened by the first data comparison unit and at least one outline pattern is greater than or equal to the contact ratio comparison threshold value or the contact ratio correction comparison threshold value;
and the second coincidence degree comparison result is that the coincidence degrees of the character outlines and all outline patterns screened by the first data comparison unit are smaller than the coincidence degree comparison threshold value or the coincidence degree correction comparison threshold value.
Specifically, the first data comparison unit calculates a definition parameter D corresponding to each text outline of the answer in the picture information according to formula (1),
in the formula (1), S represents the area of the text outline of the answer in the picture information, S0 represents the average area value of each text outline of the answer in the picture information, C represents the chromaticity value of the text outline of the answer in the picture information, and C0 represents the average chromaticity value of each text outline of the answer in the picture information.
Specifically, the first data comparison unit compares the text definition parameter D corresponding to the text outline of the answer in the picture information with a preset definition comparison parameter D1, and screens out the text outline of the answer in the picture information according to the comparison result,
if the comparison result meets a first preset condition, the first data comparison unit judges that the text outline of the answer in the picture information is screened out;
the first preset condition is that D is more than or equal to D1.
Specifically, the data processing module screens the character outline based on the definition parameters of the character outline, only judges the corresponding character of the screened character outline, in actual conditions, the character outline in the picture information uploaded by the user side has larger difference due to writing difference, and part of character outline has poorer definition or is combined with other character outlines, so that the character outline cannot be identified.
Specifically, the first analysis unit sorts the overlapping degree of the text outline screened by the first data comparison unit and each outline pattern under the first overlapping degree comparison result, and in the sorting result, the text corresponding to the outline pattern with the highest overlapping degree is used as the text corresponding to the text outline screened by the first data comparison unit.
Specifically, the second analysis unit obtains the coincidence degree judgment result of the adjacent character outlines of the character outlines screened by the first data comparison unit,
if any adjacent character outline accords with a first coincidence degree comparison result, the second analysis unit acquires characters corresponding to the adjacent character outline judged by the first analysis unit, and determines associated characters with the acquired characters based on the association relation among the characters stored in the first storage unit so as to record and generate an associated character set;
and if the adjacent character outlines all accord with the second coincidence degree comparison result, the second analysis unit judges that the characters corresponding to the character outlines screened by the first data comparison unit cannot be identified.
Specifically, according to the invention, based on the coincidence degree comparison result of the character outline and different outline patterns, different analysis units are called to carry out data processing, so that the data processing pressure is shared, when the coincidence degree of the character outline and at least one outline pattern is larger than or equal to the coincidence degree comparison threshold value or the coincidence degree correction comparison threshold value, the character outline at the moment has higher coincidence degree with the outline pattern stored in the database module, so that characters corresponding to the character outline can be judged directly based on the sorting of the coincidence degree, and the recognition efficiency of the character outline is improved.
Specifically, the data processing module of the invention indicates that the coincidence degree comparison threshold value of the character outline and each outline pattern is not high when the coincidence degree of the character outline and all outline patterns is smaller than the coincidence degree comparison threshold value or the coincidence degree correction comparison threshold value, and the associated characters of the characters corresponding to the adjacent character outline need to be determined at the moment.
Specifically, the first storage unit constructs an association relationship between each word, wherein,
the first storage unit stores a plurality of words, determines words constituting each word, and establishes association relation between each word constituting the word for any word.
Specifically, the second data comparison unit calculates a discrete parameter E according to formula (2),
in the formula (2), G (i) represents an average value of the overlapping ratio of the outline pattern corresponding to the i-th associated text in the associated text set and the outline pattern corresponding to the remaining text, n represents the number of text in the associated text set, and n is an integer greater than zero.
Specifically, the second data comparing unit receives the related text set, determines outline patterns corresponding to the related text in the related text set, compares the discrete parameter E with the discrete parameter comparison parameter E0 when the second data comparing unit obtains the coincidence ratio comparison result of the text outline screened by the first data comparing unit and the outline patterns, corrects the coincidence ratio comparison threshold H0 according to the comparison result,
the first correction mode is to correct the overlap ratio comparison threshold H0 to an overlap ratio correction comparison threshold H according to a first preset correction parameter H1, and set H=H20-H1;
the second correction mode is to correct the overlap ratio comparison threshold H0 to an overlap ratio correction comparison threshold H according to a second preset correction parameter H2, and set H=H2;
wherein h1 is smaller than h2, E is smaller than E0 in the first correction mode, and E is larger than or equal to E0 in the second correction mode.
Specifically, the preset comparison answer information is text pre-stored in a checking module, the checking module compares the text information with the preset comparison answer information, wherein,
if the text information is the same as the preset comparison answer information, the comparison module judges that the text information is correct,
and if the text information is different from the preset comparison answer information, the comparison module judges that the text information is wrong.
Specifically, the correction amount for correcting the coincidence degree contrast threshold value is determined based on the corresponding discrete parameters of the associated character set, the coincidence degree between outline patterns corresponding to all associated characters in the associated character set is higher under the condition that the discrete parameters are lower, and if the coincidence degree of any outline pattern and the outline of the characters is lower, the coincidence degree of other outline patterns and the coincidence degree of the outline of the characters are also possibly lower, so that the coincidence degree contrast threshold value needs to be reduced more, the second coincidence degree contrast result still appears after the coincidence degree contrast threshold value is prevented from being reduced, and the recognition rate of the characters corresponding to the outline of the text is improved under the premise of improving the reliability through the process.
Thus far, the technical solution of the present invention has been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of protection of the present invention is not limited to these specific embodiments. Equivalent modifications and substitutions for related technical features may be made by those skilled in the art without departing from the principles of the present invention, and such modifications and substitutions will be within the scope of the present invention.
Claims (9)
1. An image recognition-based job error question collection system, comprising:
the database module comprises a first storage unit for storing the association relation among the characters and a second storage unit for storing a plurality of character outlines, and the corresponding relation is pre-established between the character outlines and the characters;
the data acquisition module is connected with the user terminal and used for receiving the picture information sent by the user terminal;
the data processing module comprises a first data comparing unit, a second data comparing unit, a first analyzing unit and a second analyzing unit,
the first data comparison unit is connected with the data acquisition module and is used for receiving the picture information, extracting text outlines of answers in the picture information and screening the text outlines of the answers in the picture information based on definition parameters corresponding to the text outlines of the answers in the picture information;
the second data comparison unit is respectively connected with the first data comparison unit and the database module and is used for calculating the coincidence degree of the character outline screened by the first data comparison unit and each outline pattern stored in the database module, and comparing each coincidence degree with a coincidence degree comparison threshold value or a coincidence degree correction comparison threshold value so as to obtain a coincidence degree comparison result of the character outline screened by the first data comparison unit and each outline pattern;
the first analysis unit is respectively connected with the second data comparison unit and the database module, and is used for determining characters corresponding to the character outline screened by the first data comparison unit based on the coincidence degree sequencing of the character outline screened by the first data comparison unit and each outline pattern under a first coincidence degree comparison result, and generating text information according to all the determined characters;
the second analysis unit is respectively connected with the second data comparison unit and the database module and is used for determining characters corresponding to adjacent character outlines of the character outlines screened by the first data comparison unit under a second coincidence ratio comparison result, determining outline patterns corresponding to associated characters of the characters corresponding to the adjacent character outlines, and sending the outline patterns corresponding to the associated characters to the second data comparison unit so that the second data comparison unit can acquire a coincidence ratio comparison result of the character outlines screened by the first data comparison unit and the received outline patterns again after correcting the coincidence ratio comparison threshold;
the checking module is connected with the first analysis unit and used for comparing the text information generated by the first analysis unit with preset comparison answer information and judging whether the text information is wrong or not;
the first coincidence degree comparison result is that the coincidence degree of the character outline screened by the first data comparison unit and at least one outline pattern is more than or equal to the coincidence degree comparison threshold value or the coincidence degree correction comparison threshold value;
and the second coincidence degree comparison result is that the coincidence degrees of the character outlines and all outline patterns screened by the first data comparison unit are smaller than the coincidence degree comparison threshold value or the coincidence degree correction comparison threshold value.
2. The system for collecting wrong questions in operation based on image recognition according to claim 1, wherein the first data comparing unit calculates definition parameters D corresponding to each text outline of the answer in the picture information according to formula (1),
in the formula (1), S represents the area of the text outline of the answer in the picture information, S0 represents the average area value of each text outline of the answer in the picture information, C represents the chromaticity value of the text outline of the answer in the picture information, and C0 represents the average chromaticity value of each text outline of the answer in the picture information.
3. The system for collecting wrong questions in operation based on image recognition according to claim 2, wherein the first data comparing unit compares the definition parameter D corresponding to each text outline of the answer in the picture information with a preset definition comparing parameter D1, and screens out the text outline of the answer in the picture information according to the comparison result,
if the comparison result meets a first preset condition, the first data comparison unit judges that the text outline of the answer in the picture information is screened out;
the first preset condition is that D is more than or equal to D1.
4. The system for collecting wrong operation questions based on image recognition according to claim 1, wherein the first analysis unit sorts the overlapping degree of the character outline screened by the first data comparison unit and each outline pattern under the first overlapping degree comparison result, and in the sorting result, the outline pattern corresponding character with the highest overlapping degree is used as the character corresponding to the character outline screened by the first data comparison unit.
5. The system for collecting task errors based on image recognition according to claim 1, wherein the first storage unit constructs an association relationship between each text, wherein,
the first storage unit stores a plurality of words, determines words constituting each word, and establishes association relation between each word constituting the word for any word.
6. The system for collecting wrong operation questions based on image recognition according to claim 1, wherein the second analyzing unit obtains the result of determining the coincidence degree of adjacent character outlines of the character outlines screened by the first data comparing unit,
if any adjacent character outline accords with a first coincidence degree comparison result, the second analysis unit acquires characters corresponding to the adjacent character outline judged by the first analysis unit, and determines associated characters with the acquired characters based on the association relation among the characters stored in the first storage unit so as to record and generate an associated character set;
and if the adjacent character outlines all accord with the second coincidence degree comparison result, the second analysis unit judges that the characters corresponding to the character outlines screened by the first data comparison unit cannot be identified.
7. The system for collecting task errors based on image recognition according to claim 6, wherein the second data comparison unit calculates a discrete parameter E according to formula (2),
in the formula (2), G (i) represents an average value of the overlapping ratio of the outline pattern corresponding to the i-th associated text in the associated text set and the outline pattern corresponding to the remaining text, n represents the number of text in the associated text set, and n is an integer greater than zero.
8. The system for collecting wrong operation questions based on image recognition according to claim 7, wherein the second data comparing unit receives the related text set and determines outline patterns corresponding to each related text in the related text set, and when the second data comparing unit obtains the comparison result of the coincidence ratio of the text outline screened by the first data comparing unit and each outline pattern, the discrete parameter E is compared with the discrete parameter comparison parameter E0, the coincidence ratio comparison threshold H0 is corrected according to the comparison result,
the first correction mode is to correct the overlap ratio comparison threshold H0 to an overlap ratio correction comparison threshold H according to a first preset correction parameter H1, and set H=H20-H1;
the second correction mode is to correct the overlap ratio comparison threshold H0 to an overlap ratio correction comparison threshold H according to a second preset correction parameter H2, and set H=H2;
wherein h1 is smaller than h2, E is smaller than E0 in the first correction mode, and E is larger than or equal to E0 in the second correction mode.
9. The system for collecting wrong job questions based on image recognition according to claim 7, wherein the preset reference answer information is a text pre-stored in a collation module, the collation module collates the text information with the preset reference answer information, wherein,
if the text information is the same as the preset comparison answer information, the comparison module judges that the text information is correct,
and if the text information is different from the preset comparison answer information, the comparison module judges that the text information is wrong.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310102859.2A CN115841670B (en) | 2023-02-13 | 2023-02-13 | Operation wrong question collecting system based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310102859.2A CN115841670B (en) | 2023-02-13 | 2023-02-13 | Operation wrong question collecting system based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115841670A CN115841670A (en) | 2023-03-24 |
| CN115841670B true CN115841670B (en) | 2023-05-12 |
Family
ID=85579623
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310102859.2A Active CN115841670B (en) | 2023-02-13 | 2023-02-13 | Operation wrong question collecting system based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115841670B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109472014A (en) * | 2018-10-30 | 2019-03-15 | 南京红松信息技术有限公司 | A kind of wrong topic collection automatic identification generation method and its device |
| CN111104883A (en) * | 2019-12-09 | 2020-05-05 | 平安国际智慧城市科技股份有限公司 | Job answer extraction method, device, equipment and computer readable storage medium |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102982330B (en) * | 2012-11-21 | 2016-12-21 | 新浪网技术(中国)有限公司 | Character identifying method and identification device in character image |
| CN107977659B (en) * | 2016-10-25 | 2021-03-16 | 北京搜狗科技发展有限公司 | Character recognition method and device and electronic equipment |
| CN108154132A (en) * | 2018-01-10 | 2018-06-12 | 马上消费金融股份有限公司 | A kind of identity card text extraction method, system and equipment and storage medium |
| CN112287926A (en) * | 2019-07-23 | 2021-01-29 | 小船出海教育科技（北京）有限公司 | Method, device and equipment for correcting graphic questions |
| CN111242045A (en) * | 2020-01-15 | 2020-06-05 | 西安汇永软件科技有限公司 | Automatic operation exercise right and wrong indication method and system |
| CN112347997A (en) * | 2020-11-30 | 2021-02-09 | 广东国粒教育技术有限公司 | Test question detection and identification method and device, electronic equipment and medium |
| CN115393865A (en) * | 2022-08-31 | 2022-11-25 | 苏州市职业大学 | Character retrieval method, character retrieval equipment and computer-readable storage medium |
-
2023
- 2023-02-13 CN CN202310102859.2A patent/CN115841670B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109472014A (en) * | 2018-10-30 | 2019-03-15 | 南京红松信息技术有限公司 | A kind of wrong topic collection automatic identification generation method and its device |
| CN111104883A (en) * | 2019-12-09 | 2020-05-05 | 平安国际智慧城市科技股份有限公司 | Job answer extraction method, device, equipment and computer readable storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 一种面向银行票据文字自动化识别的高效人工智能方法;张振宇;姜贺云;樊明宇;;温州大学学报(自然科学版)(第03期) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115841670A (en) | 2023-03-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110738602B (en) | Image processing method and device, electronic equipment and readable storage medium | |
| US6178263B1 (en) | Method of estimating at least one run-based font attribute of a group of characters | |
| US6424728B1 (en) | Method and apparatus for verification of signatures | |
| CN108537218B (en) | Answer sheet identification processing method and device | |
| CN109460762B (en) | Answer sheet scoring method based on image recognition | |
| EP0582490A2 (en) | Converting bitmap data into page definition language commands | |
| CN106033535B (en) | Electronic paper marking method | |
| CN110689013A (en) | Automatic marking method and system based on feature recognition | |
| CN102360419A (en) | Method and system for computer scanning reading management | |
| CN107256379A (en) | Information collecting method, mobile terminal and storage medium based on image recognition | |
| CN111985465A (en) | Text recognition method, device, equipment and storage medium | |
| CN107590495A (en) | Answer sheet picture method for correcting error, device, readable storage medium storing program for executing and electronic equipment | |
| US8787702B1 (en) | Methods and apparatus for determining and/or modifying image orientation | |
| US7251380B2 (en) | Adjustment method of a machine-readable form model and a filled form scanned image thereof in the presence of distortion | |
| CN111737478A (en) | Text detection method, electronic device and computer readable medium | |
| CN113159014A (en) | Objective question reading method, device, equipment and storage medium based on handwritten question numbers | |
| CN113095307A (en) | Automatic identification method for financial voucher information | |
| CN114694161A (en) | Text recognition method and equipment for specific format certificate and storage medium | |
| CN115841670B (en) | Operation wrong question collecting system based on image recognition | |
| CN117029733B (en) | Bolt loosening detection method, system and device based on computer vision | |
| US6115506A (en) | Character recognition method, character recognition apparatus and recording medium on which a character recognition program is recorded | |
| CN104077562B (en) | A kind of scanning direction determination methods of test paper | |
| CN112926571A (en) | Formula picture recognition and question judgment system based on deep learning | |
| US20070047815A1 (en) | Image recognition apparatus, image recognition method, and image recognition program | |
| CN116049461B (en) | Question conversion system based on big data cloud platform |
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 |