CN111144191B

CN111144191B - Font identification method, font identification device, electronic equipment and storage medium

Info

Publication number: CN111144191B
Application number: CN201910750071.6A
Authority: CN
Inventors: 周林
Original assignee: Guangdong Genius Technology Co Ltd
Current assignee: Guangdong Genius Technology Co Ltd
Priority date: 2019-08-14
Filing date: 2019-08-14
Publication date: 2024-03-22
Anticipated expiration: 2039-08-14
Also published as: CN111144191A

Abstract

The embodiment of the invention discloses a font identification method, a font identification device, electronic equipment and a storage medium, wherein the method comprises the following steps: firstly, acquiring a target document image, and acquiring characters to be identified from the target document image; inputting the character to be recognized into a preset handwriting characteristic recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type; inputting the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font; and finally, determining the font type of the character to be recognized according to the first evaluation value and the second evaluation value. In this way, by constructing the feature recognition models with different font types in advance, the evaluation values of the characters belonging to different font types in the target document image can be obtained according to the different feature recognition models, so that the font types of the characters are recognized based on the evaluation values, and the font recognition of the document image is realized.

Description

Font identification method, font identification device, electronic equipment and storage medium

Technical Field

The present invention relates to the field of font processing technologies, and in particular, to a font identification method, a font identification device, an electronic device, and a storage medium.

Background

Documents typically contain two types of fonts, namely handwriting and printing. The font types of the printing fonts comprise fewer types (such as Song style, regular script, bold type and the like), the font types of the handwriting fonts comprise more types, and information can be transmitted through both fonts.

In view of the fact that a handwritten character and a print character are usually mixed in a document image, however, often the handwritten character is focused information and the print character is not focused information, and thus, the handwritten character and the print character are mixed, which results in that the handwritten character in the document image cannot be accurately processed, so how to identify the character in the document image is a problem to be solved.

Disclosure of Invention

The embodiment of the invention discloses a font identification method, a device, electronic equipment and a storage medium, which can acquire the evaluation values of all characters belonging to different font types in a target document image according to different characteristic identification models by constructing characteristic identification models of different font types in advance, so that the font types of all the characters are identified based on the evaluation values, the font identification of the document image is realized, and the problem that character processing cannot be accurately performed on a handwriting character due to the fact that handwriting fonts and printing fonts are mixed in the document image is avoided.

According to a first aspect of an embodiment of the present invention, a font identification method is disclosed, the method comprising:

acquiring a target document image, and acquiring a character to be identified from the target document image;

inputting the character to be recognized into a preset handwriting feature recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type;

inputting the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font;

and determining the font type of the character to be recognized according to the first evaluation value and the second evaluation value.

In a first aspect of the embodiment of the present invention, before the inputting the character to be recognized into the preset handwriting feature recognition model to obtain the first evaluation value of the character to be recognized belonging to the handwriting font type, the method further includes:

acquiring a handwriting character sample;

extracting handwriting characteristics from the handwriting character sample to obtain handwriting characteristics;

and constructing the handwriting characteristic recognition model through the handwriting characteristic.

In a first aspect of the embodiment of the present invention, before the inputting the character to be recognized into a preset print feature recognition model to obtain a second evaluation value that the character to be recognized belongs to a print font type, the method further includes:

Obtaining a printed character sample;

extracting the printing body characteristics of the printing character sample to obtain printing body characteristics;

and constructing the print feature recognition model through the print features.

As an optional implementation manner, in a first aspect of the embodiment of the present invention, the determining, according to the first evaluation value and the second evaluation value, a font type to which the character to be recognized belongs includes:

comparing the first evaluation value with the second evaluation value;

determining that the character to be recognized is a handwriting font under the condition that the first evaluation value is larger than the second evaluation value;

determining that the character to be recognized is a printing font under the condition that the first evaluation value is smaller than the second evaluation value;

and under the condition that the first evaluation value is equal to the second evaluation value, determining the font type of the character to be recognized according to the font recognition result of the last adjacent character of the character to be recognized and the font recognition result of the next adjacent character of the character to be recognized.

As an optional implementation manner, in the first aspect of the embodiment of the present invention, after the determining, according to the first evaluation value and the second evaluation value, a font type to which the character to be recognized belongs, the method further includes:

After all characters in the target document image are identified, dividing the target document image into at least one single test question area;

acquiring a handwriting area to be analyzed in each single test question area;

and matching the handwriting character strings in the handwriting area to be analyzed with a preset character string to obtain handwriting matching values of each single test question area.

As an optional implementation manner, in the first aspect of the embodiment of the present invention, after the matching the handwriting string in the handwriting area to be analyzed with a preset string, obtaining a handwriting matching value of each single test question area further includes:

and under the condition that the handwriting matching value of any single test question area is smaller than or equal to a matching threshold value, outputting and comparing the handwriting character string in any single test question area with the preset character string.

As an optional implementation manner, in the first aspect of the embodiment of the present invention, the acquiring a target document image includes:

receiving a target document image input by terminal equipment;

after the determining, according to the first evaluation value and the second evaluation value, the font type to which the character to be recognized belongs, the method further includes:

After all characters in the target document image are recognized, acquiring a handwritten character string to be recognized, which is included in the target document image;

and recognizing the handwriting character string to be recognized, and sending the recognition result of the handwriting character string to be recognized to the terminal equipment so that the terminal equipment can display the recognition result.

According to a second aspect of an embodiment of the present invention, there is disclosed a font recognition device, the device comprising:

the image acquisition module is used for acquiring a target document image;

the character acquisition module is used for acquiring characters to be identified from the target document image;

the first evaluation value acquisition module is used for inputting the character to be recognized into a preset handwriting characteristic recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting font type;

the second evaluation value acquisition module is used for inputting the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font;

and the font type determining module is used for determining the font type of the character to be identified according to the first evaluation value and the second evaluation value.

As an optional implementation manner, in the second aspect of the embodiment of the present invention, the apparatus further includes:

the handwritten character sample acquisition module is used for acquiring a handwritten character sample;

the handwriting feature extraction module is used for extracting handwriting features of the handwriting character samples to obtain handwriting features;

and the handwriting feature recognition model construction module is used for constructing the handwriting feature recognition model through the handwriting features.

the printed character sample acquisition module is used for acquiring a printed character sample;

the print feature extraction module is used for extracting the print features of the print character sample to obtain the print features;

and the print feature recognition model construction module is used for constructing the print feature recognition model through the print features.

As an optional implementation manner, in a second aspect of the embodiment of the present invention, the font type determining module is specifically configured to compare the first evaluation value and the second evaluation value;

the single test question area acquisition module is used for carrying out image division on the target document image to obtain at least one single test question area after all characters in the target document image are identified;

the handwriting area to be analyzed acquisition module is used for acquiring the handwriting area to be analyzed in each single test question area;

and the handwriting matching value acquisition module is used for matching the handwriting character strings in the handwriting area to be analyzed with preset character strings to acquire the handwriting matching value of each single test question area.

And the output module is used for comparing the handwriting character strings in any single test question area with the preset character strings under the condition that the handwriting matching value of any single test question area is smaller than or equal to a matching threshold value.

In a second aspect of the embodiment of the present invention, as an optional implementation manner, the image obtaining module is configured to receive a target document image input by a terminal device;

the apparatus further comprises:

the handwritten character string to be recognized is obtained by the module, and is used for obtaining the handwritten character string to be recognized included in the target document image after all characters in the target document image are recognized;

and the character string processing module is used for identifying the handwriting character string to be identified and sending the identification result of the handwriting character string to be identified to the terminal equipment so that the terminal equipment can display the identification result.

According to a third aspect of an embodiment of the present invention, there is disclosed an electronic apparatus including:

a memory storing executable program code;

a processor coupled to the memory;

the processor invokes the executable program code stored in the memory to perform the steps of the font identification method according to the first aspect of the embodiment of the present invention.

According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium storing a computer program is disclosed, wherein the computer program causes a computer to execute the font recognition method according to the first aspect of the embodiment of the present invention.

Compared with the prior art, the embodiment of the invention has the following beneficial effects:

in the embodiment of the invention, a target document image is firstly obtained, and characters to be identified are obtained from the target document image; inputting the character to be recognized into a preset handwriting feature recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type; inputting the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font; and finally, determining the font type of the character to be identified according to the first evaluation value and the second evaluation value. Therefore, through pre-constructing the feature recognition models with different font types, the evaluation values of the characters belonging to different font types in the target document image can be obtained according to the different feature recognition models, so that the font types of the characters are recognized based on the evaluation values, the font recognition of the document image is realized, and the problem that the character processing of the handwritten character cannot be accurately performed due to the fact that the handwritten character and the printing character are mixed in the document image is avoided.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of a font identification method disclosed in an embodiment of the invention;

FIG. 2 is a flow chart of another font identification method disclosed in an embodiment of the present invention;

FIG. 3 is a flow chart of another font identification method disclosed in an embodiment of the present invention;

FIG. 4 is a flow chart of another font identification method disclosed in an embodiment of the present invention;

fig. 5 is a schematic structural view of another font recognition device according to an embodiment of the present invention;

fig. 6 is a schematic structural view of another font recognition device according to an embodiment of the present invention;

fig. 7 is a schematic structural view of another font recognition device according to an embodiment of the present invention;

fig. 8 is a schematic structural view of another font recognition device according to an embodiment of the present invention;

fig. 9 is a schematic structural view of another font recognition device according to an embodiment of the present invention;

Fig. 10 is a schematic structural view of another font recognition device according to an embodiment of the present invention;

fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be noted that the terms "comprises" and "comprising," along with any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

First, an application scenario of the present invention will be described, and the present invention can be applied to a font recognition scenario, in which a document image is generally input to an electronic device so that the electronic device performs font recognition on each character in the document image. For example, if the document image is a test paper image, the test question part and the answer question part in the test paper image can be identified by the font identification method, so that the electronic equipment detects whether the answer question part is correct or not. Therefore, the font recognition method of the invention realizes accurate font recognition, so that the character processing can be performed on the handwriting characters focused by the user, and the problem that the character processing cannot be performed on the handwriting characters due to the fact that the handwriting fonts and the printing fonts are mixed in the document image is avoided.

The present invention will be described in detail with reference to specific examples.

Fig. 1 is a schematic flow chart of a font identification method disclosed in an embodiment of the present invention, where the method may be applied to an electronic device, as shown in fig. 1, and the method may include the following steps:

s101, acquiring a target document image, and acquiring characters to be recognized from the target document image.

In the embodiment of the invention, the electronic equipment can be provided with the camera, so that the target document image can be acquired through the camera; alternatively, the target document image may be acquired by a terminal device bound to the electronic device, so that the target document image is transmitted to the electronic device by the terminal device, where the terminal device may be a mobile terminal or a scanning device, etc.

The target document image may be an image corresponding to a test paper, an image corresponding to a case, an image corresponding to a bill, or the like, which is merely illustrative and the present invention is not limited thereto.

In one possible implementation, the target document image may be first subjected to image preprocessing (such as graying processing, binarizing processing, etc.); then, image segmentation is carried out on the preprocessed target document image so as to lead each character included in the preprocessed target document image to be segmented, and if the image segmentation can be realized by adopting a left-right segmentation method and an up-down segmentation method; the invention can then sequentially recognize the fonts of the characters, so that the character to be recognized can be any character in the characters.

S102, inputting the character to be recognized into a preset handwriting feature recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting font type.

In the embodiment of the present invention, the handwriting feature recognition model may be previously constructed by the following steps:

first, a handwritten character sample is acquired.

The handwriting character recognition model has the advantages that the handwriting character recognition model is high in recognition accuracy, and handwriting character samples with large data size and handwriting character samples of different users can be collected.

Then, handwriting characteristics are extracted from the handwriting Fu Yangben.

In the embodiment of the invention, the handwriting characteristics can be handwriting font proportion characteristics, handwriting font outline characteristics, handwriting font stroke density characteristics and the like.

Then, a handwriting feature recognition model is constructed from the handwriting features.

In the embodiment of the invention, the handwriting feature can be input into a first preset training model, so that the first preset training model is trained for a plurality of times, and the recognition accuracy of the trained handwriting feature recognition model is greater than or equal to a first accuracy threshold. The first preset training model may be a neural network model or the like.

S103, inputting the character to be recognized into a preset printing body characteristic recognition model to obtain a second evaluation value of the character to be recognized belonging to the printing font type.

In the embodiment of the present invention, the print feature recognition model may be previously constructed by the following steps:

first, a print character sample is obtained.

The invention can collect the print character sample with larger data quantity and collect the print character samples with different types in order to make the print character feature recognition model have higher recognition accuracy for each type of print character.

And then, extracting the printed body characteristics of the printed character sample to obtain the printed body characteristics.

In the embodiment of the invention, the print body features may be print font proportion features, print font outline features, print font stroke density features, and the like.

A print feature recognition model is then constructed from the print features.

In the embodiment of the invention, the printed body characteristics can be input into a second preset training model, so that the second preset training model is trained for a plurality of times, the recognition accuracy of the trained printed body characteristic recognition model is greater than or equal to a second accuracy threshold, and the first preset training model and the second preset training model can be the same type of model.

It should be noted that, since the printed matter may include a plurality of types such as Song Ti, regular script, bold, etc., and there are differences among the Song style, regular script, bold, etc., for example, the Song style, regular script all belong to an serif font, and bold belongs to a non serif font, the present invention may also construct a corresponding feature recognition model for different types of printed matter, so that a character to be recognized may be input into the feature recognition model corresponding to different types of printed matter, so as to obtain an evaluation value of the character to be recognized that belongs to different types of printed matter, so that in a subsequent step, a font type of the character to be recognized may be determined based on the evaluation value of the character to be recognized that belongs to different types of printed matter, and the first evaluation value of the character to be recognized that belongs to a handwriting font type.

S104, determining the font type of the character to be identified according to the first evaluation value and the second evaluation value.

In this step, the first evaluation value and the second evaluation value may be compared first, and in the case where the first evaluation value is greater than the second evaluation value, the character to be recognized is determined to be a handwriting font; under the condition that the first evaluation value is smaller than the second evaluation value, determining that the character to be recognized is a printing font; in the case that the first evaluation value is equal to the second evaluation value, the font type of the character to be recognized may be determined according to the font recognition result of the last adjacent character of the character to be recognized and the font recognition result of the next adjacent character of the character to be recognized, further, in the case that the font recognition result of the last adjacent character of the character to be recognized and the font recognition result of the next adjacent character of the character to be recognized are both handwriting fonts, the character to be recognized is determined to be handwriting fonts, in the case that the font recognition result of the last adjacent character of the character to be recognized and the font recognition result of the next adjacent character of the character to be recognized are both printing fonts, in the case that the font recognition result of the last adjacent character of the character to be recognized is different in type, the electronic device may generate a manual confirmation message, and send the manual confirmation message to the terminal device, so that the terminal device displays the character to be recognized to the user, in the case that the font recognition result of the last adjacent character of the character to be recognized is handwriting, and the font recognition result of the next adjacent character of the character to be recognized is printing font, in the case that the font recognition result of the last adjacent character of the character to be recognized is printing font, and the font recognition result of the character to be recognized is different in the font recognition result of the next adjacent character of the character to be recognized.

In order to improve the accuracy of the handwriting feature recognition model and the printing feature recognition model, the handwriting feature recognition model and the printing feature recognition model can be further trained through a manual confirmation result, for example, if the manual confirmation result of the character to be recognized is a handwriting font, the character to be recognized is input into the handwriting feature recognition model, so that the handwriting feature recognition model is trained to obtain a trained handwriting feature recognition model; if the result of the manual confirmation of the character to be recognized is the print font, inputting the character to be recognized into the print feature recognition model to train the print feature recognition model to obtain a trained print feature recognition model.

By adopting the method, firstly, a target document image is acquired, and characters to be identified are acquired from the target document image; inputting the character to be recognized into a preset handwriting characteristic recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type; inputting the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font; and finally, determining the font type of the character to be recognized according to the first evaluation value and the second evaluation value. Therefore, the characteristic difference between the handwritten font and the printed font is considered, so that the characteristic recognition models with different font types can be constructed in advance, and the evaluation values of the characters belonging to different font types in the target document image can be acquired according to the different characteristic recognition models, so that the font types of the characters are recognized based on the evaluation values, the font recognition of the document image is realized, and the problem that the character processing of the handwritten character cannot be accurately performed due to the fact that the handwritten font and the printed font are mixed in the document image is avoided.

Fig. 2 is a schematic flow chart of a font identification method disclosed in the embodiment of the present invention, as shown in fig. 2, the method may include the following steps:

s201, the electronic equipment acquires a target document image and acquires characters to be recognized from the target document image.

The specific process may refer to step S101, and will not be described in detail.

S202, the electronic equipment inputs the character to be recognized into a preset handwriting characteristic recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type.

The specific process may refer to step S102, and will not be described in detail.

S203, the electronic equipment inputs the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font.

The specific process may refer to step S103, and will not be described in detail.

S204, the electronic equipment determines the font type of the character to be recognized according to the first evaluation value and the second evaluation value.

The specific process may refer to step S104, and will not be described in detail.

S205, after all characters in the target document image are recognized, the electronic equipment divides the target document image into at least one single test question area.

Since the target document image may include a plurality of questions, and a division identifier generally exists between each question, the plurality of questions in the target document image may be divided by the division identifier to obtain a single question area corresponding to each question. Illustratively, the partition identification may be a title number, or the like, as the invention is not limited in this regard.

S206, the electronic equipment acquires the handwriting area to be analyzed in each single test question area.

In the embodiment of the invention, considering that the single test question area may include a single line of text or may include multiple lines of text, the invention can perform multi-line text detection on the single test question area first; then, under the condition that a single test question area comprises a single text, carrying out character division on the single text, namely obtaining target characters which are continuously of the same font type in the single text, determining the area where the target characters are located as character areas, wherein the character areas can be handwriting areas to be analyzed or printing body areas, namely, under the condition that the target characters are handwriting bodies, the character areas are handwriting areas to be analyzed, and under the condition that the target characters are printing bodies, the character areas are printing body areas; in the case where it is detected that a plurality of lines of text are included in a single question area, it is necessary to text-divide the plurality of lines of text to obtain a plurality of single lines of text so as to character-divide the single question area by the plurality of single lines of text. The method comprises the steps of acquiring target characters which are continuously of the same font type line by line according to the sequence of a plurality of single-line texts, determining the area where the target characters are located as a character area, and likewise, determining the character area as a handwriting area to be analyzed or a printing body area, wherein the character area is the handwriting area to be analyzed under the condition that the target characters are handwriting bodies, and determining the character area is the printing body area under the condition that the target characters are printing bodies.

Therefore, the invention can divide the characters of the single test question area to obtain at least one handwriting area to be analyzed; or alternatively; at least one handwriting area to be analyzed and at least one print area. For example, consider that the test question corresponding to the single test question area may be a selection question, an open test question, a gap-filling question, or a dictation question. Therefore, in the case that the single test question area includes an open test question, the single test question area is divided into a single printed area (i.e., an area where the open test question is located) and a single handwriting area to be analyzed (i.e., an answer area of a user for the open test question); in the case that the single test question area includes a selection question, the single test question area may be divided into a plurality of print areas (i.e., a question stem area of the selection question, an area where each item to be selected is located, etc.) and a single handwriting area to be analyzed (i.e., an option input area for the selection question by the user); in the case that the single test question area includes a blank question, the single test question area may be divided into at least one print area (i.e., at least one question stem area of the blank question) and at least one handwriting area to be analyzed (i.e., at least one text input area for the blank question by a user); it should be noted that, if the present invention is applied to a dictation scenario, that is, the user writes a document through the audio information that is listened to, the single test question area is divided into at least one handwriting area to be analyzed (that is, at least one text input area for the dictation questions) and at least one print area (that is, an area corresponding to the print existing on the dictation paper), and so on.

S207, the electronic equipment matches the handwriting character strings in the handwriting areas to be analyzed with preset character strings, and handwriting matching values of each single test question area are obtained.

In the embodiment of the present invention, the preset character string may be an answer of a test question corresponding to the preset single test question area. If the handwriting area to be analyzed in a single test question area is an option input area of a selection question, the preset character string is a standard option of the selection question, and at this time, the handwriting matching value of the single test question area can be a first numerical value (such as 100%) under the condition that the handwriting character string in the handwriting area to be analyzed is the same as the preset character string; under the condition that the handwriting character string in the handwriting area to be analyzed is different from the preset character string, the handwriting matching value of the single test question area can be a second numerical value (such as 0); if the handwriting area to be analyzed in a single test question area is an answer area of an open test question, the preset character string is a reference answer of the open test question, at this time, the invention needs to obtain a first semantic analysis result by carrying out semantic analysis on the handwriting character string in the handwriting area to be analyzed, and obtain a second semantic analysis result by carrying out semantic analysis on the preset character string, so that a handwriting matching value of the single test question area is obtained by matching the first semantic analysis result with the second semantic analysis result; if the handwriting area to be analyzed in a certain single test question area is a text input area of a blank question, the preset character string is a standard answer of the blank question, at this time, if the handwriting character string in the certain handwriting area to be analyzed is the same as the preset character string, the area matching value of the certain handwriting area to be analyzed can be a first value, and if the handwriting character string in the certain handwriting area to be analyzed is different from the preset character string, the area matching value of the certain handwriting area to be analyzed can be a second value, so that the sum value between the area matching values of all the handwriting areas to be analyzed in the certain single test question area can be calculated to obtain the handwriting matching value of the certain single test question area; if the handwriting area to be analyzed in a single test question area is a single text input area of a dictation question, the preset character string can be a standard answer of the dictation question, and at the moment, the handwriting matching value of the single test question area can be a first numerical value under the condition that the handwriting character string in the single handwriting area to be analyzed is the same as the preset character string; in the case that the handwritten character string in the single handwriting area to be analyzed is different from the preset character string, the handwritten matching value of the single test question area may be the second numerical value, and the above example is merely illustrative, and the invention is not limited thereto.

Therefore, through the steps, the electronic equipment can identify the answer parts in the single test question area, so that the answer parts are intelligently examined and approved, the examination and approval efficiency is improved, and the problem of lower examination and approval efficiency caused by manual examination and approval in the related technology is avoided.

S208, under the condition that the handwriting matching value of any single test question area is smaller than or equal to the matching threshold value, outputting and comparing the handwriting character string in any single test question area with a preset character string.

In the step, the problem that the answer of the student user is wrong or has deviation is considered, so that the standard answer and the wrong answer of the student user can be displayed together, the student user can further analyze the problem existing in the answer thought, and the learning efficiency is improved. By way of example, the handwriting character string and the preset character string can be displayed according to different character styles, namely, the handwriting character string is displayed according to a first character style, and the preset character string is displayed according to a second character style, for example, the first character style can be a number four font, a black font and the like, and the second character style can be a number three font, a thickening font, a red font and the like, so that a student user can quickly identify the character string to be checked, and the checking efficiency is improved.

Of course, in the case that the handwritten matching value of any single test question area is smaller than or equal to the matching threshold value, the invention can label the test questions corresponding to any single test question area, so that the labeled test questions can be stored in the wrong question database, and further the electronic equipment sends the test questions in the wrong question database to the terminal equipment configured by the student user according to the preset period, so that the student user can re-solve all the test questions in the wrong question database, and the complicated process of sequentially searching the wrong questions from each document is avoided. After the student user finishes solving a certain test question in the wrong question database, the new handwriting matching value of the certain test question can be re-obtained, and if the re-obtained new handwriting matching value is larger than the matching threshold value, the certain test question is deleted from the wrong question database; if the newly acquired new handwriting matching value is still smaller than or equal to the matching threshold value, keeping the certain test question stored in the wrong question database, so that the wrong question database can be updated, the current learning level of the student user is more met, and further, wrong question training is intelligently enhanced for the student user.

In an alternative embodiment of the present invention, considering that the student user makes mistakes on the test questions in a single test question area, the mistakes may be caused for the student user, or may be caused for the test question type that the student user does not fully grasp a certain test question. Based on the above, after the output comparison is completed, the error question analysis selection box can be displayed to the student user, and can comprise a first trigger control and a second trigger control, wherein the first trigger control can be a control corresponding to a test question error caused by carelessness, and the second trigger control can be a control corresponding to a test question error caused by the fact that the test question type is not mastered. In this way, under the condition that the student user triggers the first trigger control, determining that the test questions in the single test question area complete the test question analysis process; under the condition that the student user triggers the second trigger control, the test question type corresponding to the test question in the certain test question area can be obtained, a plurality of test question samples of the test question type are obtained from the question bank, and the plurality of test question samples are displayed to the student user, so that the user can practice the displayed plurality of test question samples, and understanding of the student user on the test questions of the same type is deepened. It should be noted that, since the above process is that the student user determines the cause of the wrong problem, in order to improve the intelligence, the present invention may determine the cause of the wrong problem by the following manner: after the output comparison is completed, obtaining the test question type corresponding to the test question in any single test question area, obtaining the test question of the test question type from a question library, displaying the test question to a student user, and re-obtaining the answer matching value of the student user on the test question of the test question type according to the answer content of the user on the test question, so that under the condition that the answer matching value is smaller than or equal to a matching threshold value, determining that the error question cause of the student user is possibly not mastered by the test question type; under the condition that the answer matching value is larger than the matching threshold, the problem error reason of the student user is determined to be careless, the problem error reason is not required to be determined manually, and the use experience of the user is improved.

In another optional embodiment of the present invention, considering that the learning score of the student user is usually information that the parent user is concerned about, so, in order to enable the parent user to better monitor the learning of the student user, the present invention may further obtain, after obtaining the handwriting matching value of each single test question area in the target document image, a total score of the target document image according to the handwriting matching values of all the test question areas, if the total score is less than or equal to the score threshold, send the target document image to a monitoring terminal, where the monitoring terminal is a terminal configured by the parent user, so that the parent user can timely learn the learning level of the student user; if the total score is greater than the score threshold, the target document image need not be transmitted. Therefore, the test score of the student user can be aimed at to determine whether parents are required to be notified in time, so that the parents can conduct coaching assistance and strengthen supervision force on the student user under the condition that the test score of the student user is poor.

By adopting the method, the characteristic difference between the handwritten fonts and the printed fonts is considered, so that the characteristic recognition models with different font types can be constructed in advance, the evaluation values of the characters belonging to different font types in the target document image can be obtained according to the different characteristic recognition models, the font types of the characters are recognized based on the evaluation values, the font recognition of the document image is realized, the answering part of each test question in the document image can be obtained, the intelligent examination and approval of the answering part is realized, manual examination and approval is not needed, and the examination and approval efficiency is improved.

Fig. 3 is a schematic flow chart of a font identification method disclosed in the embodiment of the present invention, as shown in fig. 3, may include the following steps:

s301, the electronic equipment acquires a target document image and acquires characters to be recognized from the target document image.

S302, the electronic equipment inputs the character to be recognized into a preset handwriting characteristic recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type.

S303, the electronic equipment inputs the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font.

S304, the electronic equipment determines the font type of the character to be recognized according to the first evaluation value and the second evaluation value.

S305, after all characters in the target document image are recognized, the electronic equipment divides the target document image into at least one single test question area.

The specific process refers to step S205, and will not be described in detail.

S306, the electronic equipment acquires the handwriting area to be analyzed in each single test question area.

The specific process may refer to step S206, and will not be described again.

S307, the electronic equipment matches the handwriting character strings in the handwriting areas to be analyzed with preset character strings, and handwriting matching values of each single test question area are obtained.

The specific process may refer to step S207, and will not be described again.

And S308, under the condition that the handwritten matching value of any one single test question area is smaller than or equal to a matching threshold value, the electronic equipment takes any one single test question area as a candidate test question area.

Under the condition that the handwritten matching value of any single test question area is smaller than or equal to a matching threshold value, the student user can be determined that the test question corresponding to any single test question area is not completely mastered, so that any single test question area can be used as a candidate test question area, and whether the student user is told about the test question corresponding to the candidate test question area in the subsequent steps is determined; under the condition that the handwriting matching value of any single test question area is larger than the matching threshold value, the student user can be determined to completely grasp the test question corresponding to any single test question area, so that the student user does not need to repeatedly explain the test question corresponding to any single test question area.

S309, the electronic equipment acquires a target test question area from the candidate test question area.

In the embodiment, the fact that a plurality of student users may not grasp the same test question in the test paper is considered to be incomplete, so that under the condition that the plurality of student users do not grasp the same test question, the invention can conduct test question explanation of the same test question for the plurality of student users, and therefore other student users who grasp the same test question do not need to spend time for subsequent video learning.

In summary, the error rate of each candidate test question area may be first obtained, where the error rate is a ratio between a first image number and a second image number, where the first image number is a number of images in which a handwriting matching value of the candidate test question area in the plurality of document images is less than or equal to a matching threshold, and the second image number is a total image number of the plurality of document images, and the plurality of document images may be test paper images of all students in one class, for example; and judging whether the error rate is greater than or equal to an error rate threshold, if the error rate is greater than or equal to the error rate threshold, determining the candidate test question area as the target test question area, and if the error rate is less than the error rate threshold, determining that the candidate test question area is not the target test question area. Thus, the candidate test question areas can be screened through the error rate of each candidate test question area, and the target test question area is obtained.

S310, the electronic equipment sends the explanation video of the test questions corresponding to the target test question area to the terminal equipment of the target student user.

In the embodiment of the invention, if the handwriting matching value of the document image in the target test question area is smaller than or equal to the matching threshold value, the student user to which the document image belongs is the target student user, so that a plurality of student users with wrong answers can be sent out aiming at the same test question. The explanation video may be a pre-recorded video for each test question.

S311, the terminal equipment of the target student user plays the explanation video.

In the embodiment of the invention, in order to improve interaction, before the terminal equipment of the target student user plays the explanation video, the target student user can be aggregated in one communication group, and the explanation video is played in the communication group, so that after the playing of the explanation video is completed, each target student user performs information interaction on the explanation video. Of course, any target student user may also send an exchange request message to the user device of the designated user through the terminal device, where the designated user is also aggregated in the exchange group in the case that the user device generates an agreement exchange request in response to the exchange request message, so that the designated user may perform information interaction with each target student user.

By adopting the method, the characteristic difference between the handwritten fonts and the printed fonts is considered, so that the characteristic recognition models with different font types can be constructed in advance, the evaluation values of the characters belonging to different font types in the target document image can be obtained according to the different characteristic recognition models, the font types of the characters are recognized based on the evaluation values, the font recognition of the document image is realized, the answering part of each test question in the document image can be obtained, the intelligent examination and approval of the answering part is realized, manual examination and approval is not needed, and the examination and approval efficiency is improved. In addition, aiming at the test questions with more error people, the invention can enable a plurality of students answering the errors to learn the error test questions in a video interpretation way, thereby improving the learning efficiency.

In another scenario of the present invention, the target document image may be a case image, and since the diagnosis instruction in the case is usually a handwriting of a doctor, the recognition difficulty of the handwriting of the doctor is high, so that the patient cannot accurately recognize the diagnosis instruction. In order to solve the problem, the present invention acquires the diagnosis instruction from the case image by the font recognition method, then recognizes the diagnosis instruction, and presents the recognition result to the patient, and the specific procedure is as described in example 4.

Fig. 4 is a schematic flow chart of a font identification method disclosed in the embodiment of the present invention, as shown in fig. 4, may include the following steps:

s401, the electronic equipment receives a target document image input by the terminal equipment, and acquires characters to be identified from the target document image.

S402, the electronic equipment inputs the character to be recognized into a preset handwriting characteristic recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type.

S403, the electronic equipment inputs the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font.

S404, the electronic equipment determines the font type of the character to be recognized according to the first evaluation value and the second evaluation value.

The font recognition method described in steps S401 to S404 may be applied to view a case scenario, in which the patient basic information on a normal case is a print font and the diagnosis description of a doctor is a handwriting font, so that the present invention may classify the print font and the handwriting font in a case image, and thus perform character recognition on the characters of the handwriting font in a subsequent step.

S405, after all characters in the target document image are recognized, the electronic equipment acquires the handwriting character strings to be recognized, which are included in the target document image.

The handwriting character string to be identified can be a diagnosis instruction of the patient.

S406, the electronic equipment identifies the handwriting character string to be identified, and sends the identification result of the handwriting character string to be identified to the terminal equipment, so that the terminal equipment displays the identification result.

In the embodiment of the invention, the handwriting character string to be recognized can be recognized through the character recognition model. Each character in the handwriting character string to be recognized is respectively input into a character recognition model to obtain a character recognition result corresponding to each character, and the character recognition results are combined according to the sequence of the characters to obtain a recognition result of the handwriting character string to be recognized. The character recognition model can be built through handwritten characters in advance, in one possible implementation manner, the electronic equipment can be provided with corresponding character recognition models aiming at different hospitals, namely, a handwritten character set of all doctors in each hospital is collected, a preset model to be trained is trained through the handwritten character set corresponding to each hospital, and the character recognition model corresponding to each hospital is obtained. Of course, the present invention may also be used to increase the accuracy of character recognition, or corresponding handwriting recognition models may be set for different doctors, and the above examples are merely illustrative, and the present invention is not limited thereto.

Further, since the target document image is a case image, a print character string corresponding to the diagnosis instruction is obtained through the character recognition model, and at this time, the terminal device can display the print character string to the user, so that the problem that the user cannot view the case due to the fact that the diagnosis instruction handwritten by the doctor is difficult to recognize is avoided. Of course, in the invention, the terminal device can play the print character string through voice, thereby avoiding the tedious operation of reading the text character string by the user.

In an optional embodiment of the present invention, the electronic device may obtain the patient's disease type according to the identification result, and send a diagnosis evaluation message to a terminal device configured by the patient, so that the patient evaluates the present visit according to the diagnosis evaluation message, and the terminal device sends the evaluation result to the electronic device, where the evaluation result may include a visit effect, a visit service, and so on. In this way, the electronic device can aggregate information of the evaluation results of the same disease type and send the evaluation results after information aggregation to the target terminal device, and the target terminal device is configured for the patient of the same disease type, so that information sharing is realized.

By adopting the method, the characteristic difference between the handwritten fonts and the printed fonts is considered, so that the characteristic recognition models with different font types can be constructed in advance, and the evaluation values of the characters belonging to different font types in the target document image can be obtained according to the different characteristic recognition models, so that the font types of the characters are recognized based on the evaluation values, the font recognition of the document image is realized, the handwriting character strings to be recognized which need character recognition can be obtained based on the font recognition result, and the character recognition can be intelligently performed on the handwriting character strings to be recognized, and the problem that the user cannot accurately recognize the diagnosis and the description in the prior art is solved.

It should be noted that, for simplicity of description, the above method embodiments are all described as a series of action combinations, but those skilled in the art should appreciate that the present invention is not limited by the described order of actions. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily required for the present invention.

Fig. 5 is a block diagram of a font recognition device according to an embodiment of the present invention, and as shown in fig. 5, may include the following modules:

An image acquisition module 501 for acquiring a target document image;

a character acquisition module 502, configured to acquire a character to be identified from the target document image;

a first evaluation value obtaining module 503, configured to input the character to be identified into a preset handwriting feature identification model, so as to obtain a first evaluation value that the character to be identified belongs to a handwriting font type;

a second evaluation value obtaining module 504, configured to input the character to be identified into a preset print feature identification model, so as to obtain a second evaluation value that the character to be identified belongs to a print font type;

and the font type determining module 505 is configured to determine, according to the first evaluation value and the second evaluation value, a font type to which the character to be recognized belongs.

Fig. 6 is a block diagram of a font recognition device according to an embodiment of the present invention, and as shown in fig. 6, may further include the following modules:

a handwritten character sample acquisition module 506, configured to acquire a handwritten character sample;

the handwriting feature extraction module 507 is configured to perform handwriting feature extraction on the handwriting character sample to obtain handwriting features;

a handwriting feature recognition model construction module 508 is configured to construct the handwriting feature recognition model from the handwriting features.

Fig. 7 is a block diagram of a font recognition device according to an embodiment of the present invention, and as shown in fig. 7, may further include the following modules:

a printed character sample acquisition module 509 for acquiring a printed character sample;

a print feature extraction module 510, configured to extract print features from the print character sample to obtain print features;

a print feature recognition model construction module 511 for constructing the print feature recognition model from the print features.

In an optional embodiment of the present invention, the font type determining module 505 is specifically configured to compare the first evaluation value with the second evaluation value;

Fig. 8 is a block diagram of a font recognition device according to an embodiment of the present invention, and as shown in fig. 8, may further include the following modules:

a single test question area obtaining module 512, configured to divide the target document image into at least one single test question area after all characters in the target document image are identified;

the handwriting area to be analyzed obtaining module 513 is configured to obtain handwriting areas to be analyzed in each of the single test question areas;

and a handwriting matching value obtaining module 514, configured to match the handwriting string in the handwriting area to be analyzed with a preset string, and obtain a handwriting matching value of each single test question area.

Fig. 9 is a block diagram of a font recognition device according to an embodiment of the present invention, and as shown in fig. 9, may further include the following modules:

and an output module 515, configured to output and compare the handwritten character string in any one of the single test question areas with the preset character string when the handwritten matching value in any one of the single test question areas is less than or equal to a matching threshold.

Fig. 10 is a block diagram of a font recognition device according to an embodiment of the present invention, and as shown in fig. 10, the image acquisition module 501 is configured to receive a target document image input by a terminal device;

The apparatus further comprises:

a handwritten character string to be recognized acquisition module 516, configured to acquire a handwritten character string to be recognized included in the target document image after all characters in the target document image are recognized;

the character string processing module 517 is configured to identify the handwriting character string to be identified, and send an identification result of the handwriting character string to be identified to the terminal device, so that the terminal device displays the identification result.

By adopting the device, firstly, a target document image is acquired, and characters to be identified are acquired from the target document image; inputting the character to be recognized into a preset handwriting characteristic recognition model to obtain a first evaluation value of the character to be recognized belonging to the handwriting type; inputting the character to be identified into a preset printing body characteristic identification model to obtain a second evaluation value of the character to be identified belonging to the type of the printing font; and finally, determining the font type of the character to be recognized according to the first evaluation value and the second evaluation value. Therefore, through pre-constructing the feature recognition models with different font types, the evaluation values of the characters belonging to different font types in the target document image can be obtained according to the different feature recognition models, so that the font types of the characters are recognized based on the evaluation values, the font recognition of the document image is realized, and the problem that the character processing of the handwritten character cannot be accurately performed due to the fact that the handwritten character and the printing character are mixed in the document image is avoided.

The specific details of the above apparatus embodiments may refer to descriptions of method embodiments, and are not repeated.

Referring to fig. 11, fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the invention. As shown in fig. 11, the electronic device may include:

a memory 1101 storing executable program code;

a processor 1102 coupled to the memory 1101;

wherein the processor 1102 invokes executable program code stored in the memory 1101 to perform some or all of the steps of the methods in the method embodiments above.

The embodiment of the invention also discloses a computer readable storage medium, wherein the computer readable storage medium stores program code, and the program code comprises instructions for executing part or all of the steps of the method in the above method embodiments.

The embodiments of the present invention also disclose a computer program product, wherein the computer program product, when run on a computer, causes the computer to perform some or all of the steps of the method as in the method embodiments above.

The embodiment of the invention also discloses an application release platform, wherein the application release platform is used for releasing a computer program product, and the computer program product is used for enabling the computer to execute part or all of the steps of the method in the method embodiments.

Those of ordinary skill in the art will appreciate that all or part of the steps of the various methods of the above embodiments may be implemented by a program that instructs associated hardware, the program may be stored in a computer readable storage medium including Read-Only Memory (ROM), random access Memory (Random Access Memory, RAM), programmable Read-Only Memory (Programmable Read-Only Memory, PROM), erasable programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), one-time programmable Read-Only Memory (OTPROM), electrically erasable programmable Read-Only Memory (EEPROM), compact disc Read-Only Memory (Compact Disc Read-Only Memory, CD-ROM) or other optical disk Memory, magnetic disk Memory, tape Memory, or any other medium that can be used for carrying or storing data that is readable by a computer.

The above describes in detail a font identification method, apparatus, electronic device and storage medium disclosed in the embodiments of the present invention, and specific examples are applied to illustrate the principles and embodiments of the present invention, where the above description of the embodiments is only for helping to understand the method and core ideas of the present invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.

Claims

1. A method of font identification, the method comprising:

comparing the first evaluation value with the second evaluation value;

2. The method according to claim 1, wherein before the character to be recognized is input into a preset handwriting feature recognition model to obtain a first evaluation value of the character to be recognized belonging to a handwriting font type, the method further comprises:

Acquiring a handwriting character sample;

3. The method according to claim 1, wherein before the inputting the character to be recognized into a preset print feature recognition model, to obtain a second evaluation value that the character to be recognized belongs to a print font type, the method further comprises:

obtaining a printed character sample;

4. A method according to any one of claims 1 to 3, wherein after said determining a font type to which said character to be recognized belongs from said first evaluation value and said second evaluation value, said method further comprises:

acquiring a handwriting area to be analyzed in each single test question area;

5. The method according to claim 4, further comprising, after the matching the handwritten character string in the handwritten area to be analyzed with a preset character string, obtaining a handwritten matching value for each of the individual question areas:

6. A method according to any one of claims 1 to 3, wherein the acquiring the target document image includes:

receiving a target document image input by terminal equipment;

7. A font recognition device, the device comprising:

the image acquisition module is used for acquiring a target document image;

the font type determining module is used for comparing the first evaluation value with the second evaluation value; determining that the character to be recognized is a handwriting font under the condition that the first evaluation value is larger than the second evaluation value; determining that the character to be recognized is a printing font under the condition that the first evaluation value is smaller than the second evaluation value; and under the condition that the first evaluation value is equal to the second evaluation value, determining the font type of the character to be recognized according to the font recognition result of the last adjacent character of the character to be recognized and the font recognition result of the next adjacent character of the character to be recognized.

8. An electronic device, the electronic device comprising:

a memory storing executable program code;

a processor coupled to the memory;

the processor invokes the executable program code stored in the memory to perform the steps of the font identification method of any of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute the steps of the font recognition method according to any one of claims 1 to 6.