Handwritten chinese character structure evaluation method, device and equipment and storage medium

Abstract

The application discloses a handwritten Chinese character structure evaluation method, device and equipment and a storage medium. The evaluation method comprises the following steps: creating a first ideographic character description sequence corresponding to a Chinese character, wherein the first ideographic character description sequence comprises component information and structure information used for describing the component information and the component structure; based on collected handwritten stroke data of the Chinese character, the features of each component in the handwritten stroke data are extracted, the features of each component are combined into structure features of the whole Chinese character, and the corresponding structure information is used as a sample label of the structure features; a structure recognition model is trained based on the structure features and the sample label; and the handwritten Chinese character to be evaluated is input into the structure recognition model for evaluation, so that an evaluation result is obtained. On the basis of the correctness determination of the component level of the written Chinese character, the Chinese character structure is further determined, higher accuracy can be provided, and therefore the application is more suitable for users with different writing styles and writing habits.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular, to a method, device, equipment and storage medium for evaluating the structure of handwritten Chinese characters. Background Art

[0002] An automatic Chinese character writing evaluation system can quickly feedback the non-standard parts of the user's writing, so as to help the user master the correct Chinese character writing norms. Existing automatic Chinese character writing evaluation systems usually evaluate from the perspective of whether the strokes are written correctly, that is, whether the stroke types are written correctly, whether the stroke numbers are correct, whether the stroke orders are correct, etc.; and evaluate from the perspective of whether the components are written correctly, that is, whether the component types are correct, whether the component numbers are correct, etc. However, the stroke orders and component compositions of different Chinese characters may be exactly the same. For example, for the Chinese characters "ling" and "le", their stroke orders and component compositions are exactly the same. Therefore, it is not sufficient to assert that the Chinese character writing is correct just because the strokes and components are written correctly. Summary of the Invention

[0003] In view of the above technical problems, the present invention provides a method, device, equipment and storage medium for evaluating the structure of handwritten Chinese characters. After corresponding verification of the strokes and components of the written Chinese characters, the method can further verify the correctness of Chinese character writing from the perspective of Chinese character structure, so that users can master more detailed and complete Chinese character writing norms.

[0004] Other features and advantages of the present disclosure will become apparent from the following detailed description, or be learned in part through the practice of the present disclosure.

[0005] According to one aspect of the present invention, a method for evaluating the structure of handwritten Chinese characters is provided. The evaluation method includes:

[0006] Create a first ideographic character description sequence corresponding to the Chinese character. The first ideographic character description sequence includes component information and structure information for describing the composition structure of the component information;

[0007] Based on the collected handwritten stroke data of Chinese characters, extract the features of each component in the handwritten stroke data, and form the features of each component into the structure features of the whole Chinese character, and use the corresponding structure information as the sample label of the structure features;

[0008] Train a structure recognition model based on the structure features and the sample labels;

[0009] Input the handwritten Chinese character to be evaluated into the structure recognition model for evaluation to obtain an evaluation result.

[0010] Further, when creating the first ideographic character description sequence corresponding to the Chinese character, it further includes:

[0011] Create the first ideographic character description sequence corresponding to all Chinese characters to be evaluated, and construct a query table to obtain all the first ideographic character description sequences;

[0012] When using the corresponding structural information as the sample label of the structural feature, the specific methods include:

[0013] Extract the structural information from the first semantic text description sequence corresponding to the query table, and use the extracted structural information as the sample label of the structural feature.

[0014] Furthermore, the structural information includes one or more of the following:

[0015] Left-right structure; top-bottom structure; left-middle-right structure; top-middle-bottom structure; top-enclosed structure; bottom-enclosed structure; left-enclosed structure; fully enclosed structure; inlaid structure.

[0016] Furthermore, the step of extracting features from each component of the handwritten handwriting data and combining the features of each component to form the structural features of the entire Chinese character includes:

[0017] The trajectory point sequence of each stroke in the handwritten handwriting data of Chinese characters is divided into the corresponding components according to the relationship between stroke and component;

[0018] Extract the coordinate features and label features of each component, and concatenate the coordinate features and label features to obtain the feature vectors corresponding to each component.

[0019] The structural features of the entire Chinese character are composed of the feature vectors of all its components.

[0020] Furthermore, the handwritten handwriting data includes a sequence of trajectory points of the strokes that make up Chinese characters. Each trajectory point of each stroke has coordinates. The coordinate features are composed of the coordinates of the lower left and upper right points of the circumscribed rectangle formed by all the trajectory points contained in the component, as well as the centroid coordinates. The marking features are used to indicate the existence of the trajectory point region in the circumscribed rectangle of the component.

[0021] Furthermore, the structure recognition model is obtained based on a sequence-to-sequence model structure.

[0022] Furthermore, the step of inputting the handwritten Chinese character to be evaluated into the structure recognition model for evaluation, and obtaining the evaluation result, includes:

[0023] The components written by the handwritten Chinese characters are determined;

[0024] When the determination result of the component written by the handwritten Chinese character is correct, the structural features of the trajectory point sequence of the handwritten Chinese character are extracted;

[0025] Based on the extracted structural features and the structural recognition model, the structural information of the handwritten Chinese characters is predicted;

[0026] The prediction results of the structural information are combined with the written components of the handwritten Chinese character to obtain the second ideographic character description sequence of the current handwritten Chinese character;

[0027] The evaluation result is obtained by comparing the second ideographic character description sequence with the correct first ideographic character description sequence corresponding to the handwritten Chinese character.

[0028] According to a second aspect of this disclosure, a handwritten Chinese character structure evaluation device is provided, comprising: an information creation module for creating a first ideographic character description sequence corresponding to a Chinese character, the first ideographic character description sequence including component information and structural information describing the compositional structure of the component information; a feature extraction module for extracting features of each component in the collected handwritten handwriting data of Chinese characters, combining the features of each component into the structural features of the entire Chinese character, and using the corresponding structural information as sample labels for the structural features; a model generation module for training a structure recognition model based on the structural features and the sample labels; and an evaluation module for inputting the handwritten Chinese character to be evaluated into the structure recognition model for evaluation, and obtaining an evaluation result.

[0029] According to a third aspect of this disclosure, a handwritten Chinese character structure evaluation device is provided, comprising: a processor; and a memory arranged to store computer-executable instructions, wherein the executable instructions, when executed, cause the processor to: create a first ideographic character description sequence corresponding to a Chinese character, the first ideographic character description sequence including component information and structural information for describing the compositional structure of the component information; extract features of each component in the handwritten handwriting data based on collected handwritten handwriting data of Chinese characters, and combine the features of each component into structural features of the entire Chinese character, and use the corresponding structural information as sample labels for the structural features; train a structural recognition model based on the structural features and the sample labels; input the handwritten Chinese character to be evaluated into the structural recognition model for evaluation, and obtain an evaluation result.

[0030] According to a fourth aspect of this disclosure, a computer-readable storage medium is provided that stores a computer program, which, when executed by a processor, performs the evaluation method described above.

[0031] The technical solution disclosed herein has the following beneficial effects:

[0032] 1. Based on the correctness judgment of the component level of the written Chinese characters, the further judgment of the Chinese character structure can provide users with more detailed and complete feedback on the correctness of the written Chinese characters.

[0033] 2. When predicting the structure of Chinese characters, only the abstract features of the strokes of the components are used, so that the model only uses the position and shape information of the components when predicting the structure of Chinese characters, thereby improving the generalization ability of the model.

[0034] 3. It adopts a model-based evaluation method, which can handle Chinese character structure prediction in various special scenarios and provide higher accuracy, making it more suitable for users with different writing styles and habits. Attached Figure Description

[0035] Figure 1 This is a flowchart of a handwritten Chinese character structure evaluation method as described in the embodiments of this specification;

[0036] Figure 2 This is a schematic diagram illustrating exemplary handwritten Chinese characters in the embodiments of this specification;

[0037] Figure 3 A flowchart of the method for evaluating handwritten Chinese characters using a structural recognition model in the embodiments of the specification;

[0038] Figure 4 Example diagrams illustrating the evaluation of correct writing of Chinese character structures in the embodiments of the instruction manual;

[0039] Figure 5 Example diagrams illustrating the evaluation of incorrect Chinese character structure writing in the embodiments of the instruction manual;

[0040] Figure 6 This is a schematic diagram of a handwritten Chinese character structure evaluation device in an embodiment of this specification;

[0041] Figure 7 This is a terminal device used to implement a method for evaluating the structure of handwritten Chinese characters, as described in the embodiments of this specification.

[0042] Figure 8 This is a computer-readable storage medium used in the embodiments of this specification to implement a method for evaluating the structure of handwritten Chinese characters. Detailed Implementation

[0043] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more of the specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0044] Furthermore, the accompanying drawings are merely illustrative of this disclosure. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0045] like Figure 1 As shown in the embodiments of this specification, a method for evaluating the structure of handwritten Chinese characters is provided. The execution subject of this method can be a terminal device, such as a mobile phone, tablet computer, or personal computer. The method specifically includes the following steps S101 to S104:

[0046] In step S101, a first ideographic character description sequence corresponding to the Chinese character is created. The first ideographic character description sequence includes component information and structural information for describing the composition structure of the component information.

[0047] Chinese characters, as ideographic writing systems, represent meaning through combinations of different shapes. These shapes typically refer to components. Based on the number of components, Chinese characters can be divided into single-component characters and compound characters. Single-component characters are those with a single component, while compound characters are composed of multiple components combined through a specific structure. In compound characters, any Chinese character can be described by specifically combining components and structures. The sequence obtained through these specific combinations is usually called the ideographic character description sequence, and different Chinese characters have different ideographic character description sequences. Structural information refers to the information describing the component structure of a Chinese character; component information refers to the components contained within the character.

[0048] Exemplary structural information can be divided into: left-right structure, with Indicates; upper-lower structure, with Indicates; left-middle-right structure, with Indicates; upper-middle-lower structure, with Indicates; upper-right surrounded structure, with Indicates; upper-left surrounded structure, with Indicates; lower-left surrounded structure, with Indicates; upper surrounded structure, with Indicates; lower surrounded structure, with Indicates; left surrounded structure, with Indicates; full surrounded structure, with Indicates; inlaid structure, with Indicates.

[0049] Since the first ideographic character description sequences of different Chinese characters should be different, the constructed first ideographic character description sequence query table is used as the basis for judging whether the Chinese character structure is correct. For example, for the Chinese character "结", its component information can be expressed as "纟士口". According to its component composition structure, that is, the structure information, the first ideographic character description sequence of "结" can be expressed as

[0050] In step S102, based on the collected handwritten stroke data of Chinese characters, extract the features of each component in the handwritten stroke data, and form the structural features of the entire Chinese character from the features of each component, and use the corresponding structure information as the sample label of the structural features.

[0051] Among them, when creating the first ideographic character description sequence corresponding to a Chinese character, it is also necessary to collect sufficient training samples. The training samples are composed of the handwritten strokes of Chinese characters. To make these handwritten strokes analyzable, these handwritten strokes are digitized, and the features of relevant components are extracted, so that the handwritten Chinese characters have analyzable structural features. And since they are training samples, the structure information of the standard Chinese characters corresponding to these handwritten Chinese characters is known. Therefore, the known structure information is used as the sample label of the handwritten Chinese characters, so that the corresponding model can be trained.

[0052] It should be noted that in this step, it is possible to choose to remove the component information in the first ideographic character description sequence of the Chinese character, leaving only the structure information and using the structure information as the label of the current sample.

[0053] In step S103, based on the structural features and the sample labels, train a structure recognition model. <000……​​​​Since the training samples of the structure recognition model are the structural features and corresponding structural information of handwritten Chinese characters, when analyzing the Chinese characters to be evaluated, the structure recognition model can infer the structural information from the structural features of the handwritten Chinese characters to be evaluated, thereby further evaluating the correctness of Chinese character writing from the perspective of Chinese character structure.

[0056] In one embodiment, step S101 further includes:

[0057] Create the first ideographic character description sequence corresponding to all Chinese characters to be evaluated, and construct a query table for all the first ideographic character description sequences.

[0058] Since the construction of the first ideographic character description sequence typically involves all Chinese characters or commonly used Chinese characters, building a lookup table allows for direct retrieval of the corresponding first ideographic character description sequences during subsequent model training and character evaluation, thereby improving efficiency. Especially during model training, structural information can be directly extracted from the corresponding first ideographic character description sequences in the lookup table, enabling the extracted structural information to be used as sample labels for structural features.

[0059] In one embodiment, in step S102, extracting the features of each component in the handwritten handwriting data and combining the features of each component to form the structural features of the entire Chinese character includes:

[0060] The trajectory point sequence of each stroke in the handwritten handwriting data of Chinese characters is divided into corresponding components according to the relationship between stroke and component. The coordinate features and the label features of each component are extracted, and the coordinate features and the label features are concatenated to obtain the feature vector corresponding to each component. Based on the feature vectors of all components, the structural features of the entire Chinese character are formed.

[0061] As a supplement, the handwritten handwriting data includes a sequence of trajectory points of the strokes that make up Chinese characters. Each trajectory point of each stroke has coordinates. The coordinate features are composed of the coordinates of the lower left and upper right points of the circumscribed rectangle formed by all the trajectory points contained in the component, as well as the centroid coordinates. The marking features are used to indicate the existence of the trajectory point region in the circumscribed rectangle of the component.

[0062] The following will provide an illustrative explanation of the principles for obtaining training samples and the training principles of the model.

[0063] In handwriting data, when a Chinese character has n strokes, the number of trajectory points of the i-th stroke during writing is ci. The handwriting data is represented as S={(x 11 ,y 11 ),(x 12 ,y12 ),...,(x 1c1 ,y 1c1 ),...,(x ncn ,y ncn )}, where (x ij ,y ij ) represents the coordinates of the j-th trajectory point of the i-th stroke of a Chinese character.

[0064] Based on this formula S, handwritten handwriting data can be collected and analyzed, and thus used as training samples for building a Chinese character structure recognition model. The construction process of the structure recognition model includes dividing the trajectory point sequence of each stroke in the handwritten handwriting data into corresponding components according to the relationship between strokes and components. Then, corresponding features are extracted from each component, and the features extracted from each component are combined to form the features of the Chinese character. The specific steps for extracting features from each component are as follows:

[0065] Extract the coordinate features formed by the lower left and upper right coordinates and the centroid coordinates of the bounding rectangle containing all trajectory points of the component. Assume the component has m trajectory points, and its trajectory point sequence can be represented as R. i ={(x1,y1),...,(x m ,y m The minimum values ​​of these m trajectory points on the x-axis and y-axis are respectively x min and y min The maximum values ​​are x max and y max .

[0066] The average value of m trajectory points on the x-axis is expressed as:

[0067]

[0068] The average value of m trajectory points on the y-axis is expressed as:

[0069]

[0070] The lower left coordinate of the circumscribed rectangle can be represented as (x min ,y min The lower right coordinate can be represented as (x max ,y max The centroid coordinates can be represented as (x... mean ,y mean The coordinate characteristics can be represented as [x min ,y min ,x max ,y max ,x mean ,y mean ].

[0071] The marking features need to be extracted so that the existence of the trajectory point area in the external rectangle of the component can be indicated. According to the preset parameter grid, the external rectangle of the component is divided into grid*grid regions of equal size, and then it is sequentially detected whether there are trajectory points belonging to the component in these grid*grid regions. If there are, it is marked as 1, otherwise it is marked as 0. Finally, a marking feature vector [g1, g2,..., g grid*grid is obtained, where g i ∈ {0, 1}.

[0072] The coordinate feature vector and the marking feature vector are concatenated to obtain the feature vector of the component with a length of L:

[0073] feature = [x min , y min , x max , y max , x mcan , y mean , g1, g2,..., g grid*grid

[0074] where L = 6 + grid*grid.

[0075] Remove the component information in the ideographic description sequence of Chinese characters and only leave the structural information, and use the structural information as the label of the current sample.

[0076] Specifically, as Figure 2 shown, given the Chinese character "结" and its handwritten handwriting data, and constructing the external rectangle of the components it contains, its corresponding ideographic description sequence can be expressed as Then the true label of its structure prediction is where "*" is a component placeholder, used to represent any component, so that the model only needs to consider the Chinese character structure during prediction without considering the structure composed of which components. After the above steps, the training sample corresponding to the Chinese character "结" can be expressed as {feature1, feature2, feature3, label}, where feature i represents the feature vector extracted from the i-th component,

[0077] In one embodiment, the structure recognition model is obtained based on the sequence-to-sequence model structure.

[0078] ​Among them, a Chinese character structure recognition model is constructed based on the sequence-to-sequence model structure, and the constructed training samples are used for training. The sequence-to-sequence model usually consists of two parts, namely an encoder and a decoder. The encoder is used to receive input features for deep feature extraction, and the decoder is used to receive the deep features extracted by the encoder to predict the output sequence. When constructing the Chinese character structure recognition model, the encoder can adopt a multi-layer BiLSTM, that is, a bidirectional long short-term memory network, which is used to receive Chinese character structure features and perform deep feature extraction. The decoder can adopt a multi-layer GRU combined with an attention mechanism. GRU is a gated recurrent unit model, which is used to receive the deep features extracted by the encoder to predict the structure sequence.

[0079] In one embodiment, after obtaining the structure recognition model, the correctness evaluation of Chinese character writing can be carried out. As Figure 3 shown, the evaluation process includes steps S301 - S305:

[0080] In step S301, the components written in the handwritten Chinese character are judged.

[0081] In step S302, when the judgment result of the components written in the handwritten Chinese character is correct, the structural features of the trajectory point sequence of the handwritten Chinese character are extracted.

[0082] In step S303, according to the extracted structural features and the structure recognition model, the structural information of the handwritten Chinese character is predicted.

[0083] In step S304, the prediction result of the structural information is combined with the components written in the handwritten Chinese character to obtain the second ideographic character description sequence of the current handwritten Chinese character.

[0084] In step S305, the second ideographic character description sequence is compared with the correct first ideographic character description sequence corresponding to the handwritten Chinese character to obtain the evaluation result.

[0085] Among them, as Figure 4 shown, an evaluation example when the Chinese character structure is correctly written is provided. Figure 4 It is an example of the correct writing of the Chinese character structure of the Chinese character "qian". The components of the Chinese character "qian" are Its correct first ideographic character description sequence in the ideographic Chinese character description sequence query table is Using the Chinese character structure recognition model for Figure 4 the corresponding written Chinese character to perform the correctness evaluation of Chinese character writing, the obtained second ideographic character description sequence is Therefore, it can be determined that the Chinese character structure of the current written Chinese character is correct.

[0086] As Figure 5 shown Figure 5 is an example of incorrect writing of the Chinese character structure of the Chinese character "qian". Using the Chinese character structure recognition model to perform a correctness evaluation of the written Chinese character corresponding to Figure 5 results in a second semantic Chinese character description sequence as which does not conform to the correct semantic Chinese character description sequence. Therefore, it can be determined that the Chinese character structure of the current written Chinese character is incorrect, and the current written Chinese character "qian" has been written from the original upper-lower structure to the left-right structure.

[0087] Based on the same idea, as Figure 6 shown, an exemplary embodiment of the present disclosure also provides a handwritten Chinese character structure evaluation device 600, including an information creation module 601 for creating a first semantic Chinese character description sequence corresponding to a Chinese character, the first semantic Chinese character description sequence including component information and structure information for describing the composition structure of the component information; a feature extraction module 602 for extracting the features of each component in the handwritten handwriting data based on the collected handwritten handwriting data of the Chinese character, and forming the features of the entire Chinese character structure from the features of each component, and using the corresponding structure information as the sample label of the structure feature; a model generation module 603 for training a structure recognition model based on the structure feature and the sample label; an evaluation module 604 for inputting the handwritten Chinese character to be evaluated into the structure recognition model for evaluation to obtain an evaluation result.

[0088] Using the above handwritten Chinese character structure evaluation device 600, on the basis of determining the correctness of the written Chinese character at the component level, a further Chinese character structure determination can give the user a more detailed and complete feedback on the correctness of the Chinese character writing; when predicting the Chinese character structure, only the abstract features of the component handwriting are used, so that the model only uses the position and shape information of the components when predicting the Chinese character structure, thereby improving the generalization ability of the model; adopting a model-based evaluation method can handle the Chinese character structure prediction in various special scenarios and provide a higher accuracy rate, so it is more suitable for users with different writing styles and writing habits.

[0089] The specific details of each module / unit in the above device have been described in detail in the embodiments of the method part, and the undisclosed detailed content can be seen in the content of the embodiments of the method part, so it will not be repeated here.

[0090] Based on the same idea, the embodiments of this specification also provide a handwritten Chinese character structure evaluation device, as Figure 7 shown.

[0091] The handwritten Chinese character structure evaluation device can be the terminal device or server provided in the above embodiments.

[0092] Handwritten Chinese character structure evaluation devices can vary considerably due to differences in configuration or performance. They may include one or more processors 701 and memory 702, with memory 702 storing one or more application programs or data. Memory 702 may include readable media in the form of volatile memory cells, such as random access memory (RAM) and / or cache memory, and may further include read-only memory cells. The application programs stored in memory 702 may include one or more program modules (not shown in the figures), such program modules including, but not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Furthermore, processor 701 may be configured to communicate with memory 702 and execute a series of computer-executable instructions stored in memory 702 on the handwritten Chinese character structure evaluation device. The handwritten Chinese character structure evaluation device may also include one or more power supplies 703, one or more wired or wireless network interfaces 704, one or more I / O interfaces (input / output interfaces) 705, and one or more external devices 706 (e.g., keyboards, drawing tablets, Bluetooth devices, etc.). It can also communicate with one or more devices that enable user interaction with the device, and / or with any device that enables the device to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication can be performed through the I / O interface 705. Furthermore, the device can also communicate with one or more networks (e.g., local area networks (LANs)) through the wired or wireless interface 704.

[0093] Specifically, in this embodiment, the handwritten Chinese character structure evaluation device includes a memory and one or more programs, wherein one or more programs are stored in the memory, and one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the handwritten Chinese character structure evaluation device, and is configured to be executed by one or more processors. The one or more programs include computer-executable instructions for performing the following:

[0094] A first ideographic character description sequence corresponding to a Chinese character is created. The first ideographic character description sequence includes component information and structural information for describing the compositional structure of the component information. Based on collected handwritten handwriting data of Chinese characters, features of each component in the handwritten handwriting data are extracted, and the features of each component are combined to form the structural features of the entire Chinese character. The corresponding structural information is used as sample labels for the structural features. Based on the structural features and the sample labels, a structural recognition model is trained. The handwritten Chinese character to be evaluated is input into the structural recognition model for evaluation, and the evaluation result is obtained.

[0095] Based on the same idea, exemplary embodiments of this disclosure also provide a computer-readable storage medium storing a program product capable of implementing the methods described above. In some possible embodiments, various aspects of this disclosure can also be implemented as a program product including program code, which, when the program product is run on a terminal device, causes the terminal device to perform the steps described in the "Exemplary Methods" section of this specification according to the various exemplary embodiments of this disclosure.

[0096] refer to Figure 8 As shown, a program product 800 for implementing the above-described method according to an exemplary embodiment of the present disclosure is described. This product may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, the readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0097] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0098] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0099] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0100] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing devices can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0101] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the method according to the exemplary embodiments of this disclosure.

[0102] Furthermore, the above figures are merely illustrative representations of the processes included in the methods according to exemplary embodiments of this disclosure, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0103] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to exemplary embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0104] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0105] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for evaluating the structure of handwritten Chinese characters, characterized in that, The evaluation methods include: Create a first ideographic character description sequence corresponding to a Chinese character, the first ideographic character description sequence including component information and structural information for describing the compositional structure of the component information; Based on the collected handwritten handwriting data of Chinese characters, the features of each component in the handwritten handwriting data are extracted, and the features of each component are combined to form the structural features of the entire Chinese character, and the corresponding structural information is used as the sample label of the structural features. Based on the structural features and the sample labels, a structural recognition model is trained. The handwritten Chinese character to be evaluated is input into the structure recognition model for evaluation, and the evaluation result is obtained. Specifically, this includes: determining the components written by the handwritten Chinese character; when the determination result of the components written by the handwritten Chinese character is correct, extracting the structural features of the trajectory point sequence of the handwritten Chinese character; predicting the structural information of the handwritten Chinese character based on the extracted structural features and the structure recognition model; combining the predicted result of the structural information with the components written by the handwritten Chinese character to obtain the second ideographic character description sequence of the current handwritten Chinese character; comparing the second ideographic character description sequence with the correct first ideographic character description sequence corresponding to the handwritten Chinese character to obtain the evaluation result.

2. The method for evaluating the structure of handwritten Chinese characters according to claim 1, characterized in that, When creating the first ideographic character description sequence corresponding to the Chinese character, the following is also included: Create the first ideographic character description sequence corresponding to all Chinese characters to be evaluated, and construct a query table to obtain all the first ideographic character description sequences; When using the corresponding structural information as the sample label for the structural feature, the specific steps include: Extract the structural information from the first semantic text description sequence corresponding to the query table, and use the extracted structural information as the sample label of the structural feature.

3. The method for evaluating the structure of handwritten Chinese characters according to claim 1, characterized in that, The structural information includes one or more of the following: Left-right structure; top-bottom structure; left-middle-right structure; top-middle-bottom structure; top-enclosed structure; bottom-enclosed structure; left-enclosed structure; fully enclosed structure; inlaid structure.

4. The method for evaluating the structure of handwritten Chinese characters according to claim 1, characterized in that, The step of extracting features from each component of the handwritten handwriting data and combining the features of each component to form the structural features of the entire Chinese character includes: The trajectory point sequence of each stroke in the handwritten handwriting data of Chinese characters is divided into the corresponding components according to the relationship between stroke and component; Extract the coordinate features and label features of each component, and concatenate the coordinate features and label features to obtain the feature vectors corresponding to each component. The structural features of the entire Chinese character are composed of the feature vectors of all its components.

5. The method for evaluating the structure of handwritten Chinese characters according to claim 4, characterized in that, The handwritten handwriting data includes a sequence of trajectory points of the strokes that make up Chinese characters. Each trajectory point of each stroke has coordinates. The coordinate features are composed of the coordinates of the lower left and upper right points of the outer rectangle formed by all the trajectory points contained in the component, as well as the centroid coordinates. The marking features are used to indicate the existence of the trajectory point region in the outer rectangle of the component.

6. The method for evaluating the structure of handwritten Chinese characters according to claim 1, characterized in that, The structure recognition model is derived from a sequence-to-sequence model structure.

7. A device for evaluating the structure of handwritten Chinese characters, comprising: The information creation module is used to create a first ideographic character description sequence corresponding to a Chinese character. The first ideographic character description sequence includes component information and structural information for describing the compositional structure of the component information. The feature extraction module is used to extract the features of each component in the collected handwritten handwriting data of Chinese characters, combine the features of each component into the structural features of the whole Chinese character, and use the corresponding structural information as the sample label of the structural features. The model generation module is used to train a structure recognition model based on the structural features and the sample labels; The evaluation module is used to input the handwritten Chinese character to be evaluated into the structure recognition model for evaluation and obtain the evaluation result. Specifically, it includes: determining the components written by the handwritten Chinese character; when the determination result of the components written by the handwritten Chinese character is correct, extracting the structural features of the trajectory point sequence of the handwritten Chinese character; predicting the structural information of the handwritten Chinese character based on the extracted structural features and the structure recognition model; combining the predicted result of the structural information with the components written by the handwritten Chinese character to obtain the second ideographic character description sequence of the current handwritten Chinese character; and comparing the second ideographic character description sequence with the correct first ideographic character description sequence corresponding to the handwritten Chinese character to obtain the evaluation result.

8. A handwritten Chinese character structure evaluation device, comprising: processor; and a memory configured to store computer-executable instructions, which, when executed, cause the processor to: Create a first ideographic character description sequence corresponding to a Chinese character, the first ideographic character description sequence including component information and structural information for describing the compositional structure of the component information; Based on the collected handwritten handwriting data of Chinese characters, the features of each component in the handwritten handwriting data are extracted, and the features of each component are combined to form the structural features of the entire Chinese character, and the corresponding structural information is used as the sample label of the structural features. Based on the structural features and the sample labels, a structural recognition model is trained. The handwritten Chinese character to be evaluated is input into the structure recognition model for evaluation, and the evaluation result is obtained. Specifically, this includes: determining the components written by the handwritten Chinese character; when the determination result of the components written by the handwritten Chinese character is correct, extracting the structural features of the trajectory point sequence of the handwritten Chinese character; predicting the structural information of the handwritten Chinese character based on the extracted structural features and the structure recognition model; combining the predicted result of the structural information with the components written by the handwritten Chinese character to obtain the second ideographic character description sequence of the current handwritten Chinese character; comparing the second ideographic character description sequence with the correct first ideographic character description sequence corresponding to the handwritten Chinese character to obtain the evaluation result.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the handwritten Chinese character structure evaluation method as described in any one of claims 1 to 6.

Images