Chinese character writing commenting method and system

Abstract

The invention provides a Chinese character writing commenting method and system. The method belongs to the technical field of Chinese character writing comment, and comprises the following steps: obtaining a Chinese character image to be commented, and extracting a single Chinese character image; preprocessing any single Chinese character image to obtain a primary processing image; using an edge detection algorithm to extract contours of Chinese characters in the primary processing image; and judging the font type of the Chinese character based on the contour of the Chinese character, inputting the Chinese character into a corresponding preset Chinese character comment model according to the font type, analyzing the radical, the radical and the overall structure of the Chinese character to obtain a comment result, and displaying the comment result. The system comprises an acquisition module, a preprocessing module, an extraction module and a display module. Through the commenting method and system, the accuracy of Chinese character writing commenting can be improved.

Description

Technical Field

[0001] The present application relates to the field of Chinese character review, and more specifically, to a method and system for reviewing Chinese character writing. Background Art

[0002] In the prior art, the comments on Chinese character writing are generally done manually, which is time-consuming and laborious, or the comments are template-based and lack specific targeting. Summary of the invention

[0003] The purpose of this application is to provide a method and system for commenting on Chinese character writing, which can increase the accuracy of Chinese character writing comments.

[0004] This application is implemented as follows: In a first aspect, the present application provides a method for reviewing Chinese character writing, comprising: The Chinese character images to be reviewed are obtained, and a single Chinese character image is extracted therefrom; any single Chinese character image is preprocessed to obtain a preliminary processed image; the outline of the Chinese character in the preliminary processed image is extracted using an edge detection algorithm; the font type of the Chinese character is determined based on the outline of the Chinese character, and the Chinese character is input into the corresponding pre-set Chinese character review model according to the font type to analyze the radical, component and overall structure of the Chinese character to obtain the review result and display it.

[0005] Based on the first aspect, the step of obtaining the Chinese character image to be commented on specifically includes: Create a Chinese character review application, and add network, camera, and storage permissions to the created project file; the review interface of the Chinese character review application has an image upload icon and a photo taking icon; When it is detected that the image upload icon is selected, an image upload request is generated, and the image selection interface is jumped to according to the image upload request. At the same time, the network permission is checked. If it passes, the selected image is marked as the Chinese character image to be reviewed and uploaded to the application; When it is detected that the photo icon is selected, a photo request is generated and the network permission is checked. If it is passed, the camera is opened to take a photo, and the image taken this time is transmitted to the image display interface. After the photo is completed, it jumps to the image display interface to wait for selection. After the selection is confirmed, it will be marked as a Chinese character image to be commented on and uploaded to the application.

[0006] Based on the first aspect, the specific steps of extracting a single Chinese character image include: Use OpenCV's contour detection algorithm to find all contours in the image; According to the size and shape of the outline, the outline representing the Chinese character is selected; According to the filtered contours, the image is segmented into individual Chinese character images.

[0007] Based on the first aspect, the specific steps of preprocessing any single Chinese character image to obtain a preliminary processed image include: Process any single Chinese character image and convert it into a grayscale image; Convert grayscale image to binary image; A filtering algorithm is used to remove noise from the binary image to obtain a preliminary processed image.

[0008] Based on the first aspect, the specific steps of removing noise in the binary image using the filtering algorithm include: Determine the noise type in the binary image, the noise type including one or more of salt and pepper noise, speckle noise, Gaussian noise, small noise points and holes; For salt and pepper noise and speckle noise, select the corresponding filter according to the size of the image and replace the pixel value with the median of the pixel values ​​in its neighborhood for each pixel in the image; For Gaussian noise, the corresponding filter is selected according to the size of the image. For each pixel in the image, the pixel value is replaced by the average value of the pixel values ​​in its neighborhood. For small noise points and holes, the image is eroded to remove the small noise points, and the image is expanded to fill the holes.

[0009] Based on the first aspect, the specific steps of extracting the contours of Chinese characters in the preliminary processed image using the edge detection algorithm include: Use the function cv2.Gaussian Blur(gray, (5, 5), 0) to blur the image, where (5, 5) is the size of the blur kernel. Use the function cv2.Canny(gray, 75, 200) to perform Canny edge detection to obtain the edge detection image, where 75 and 200 are threshold parameters; Use cv2.findContours() function to find contours in edge detection image; Filter out the outline of Chinese characters according to the area and shape characteristics of the outline; Use cv2.contourArea() function to calculate the area of ​​the contour, and use conditional statements to filter to get the filtered image; Draw the filtered contours on the filtered image using cv2.drawContours() function.

[0010] Based on the first aspect, the specific steps of determining the font type of a Chinese character based on the outline of the Chinese character include: Divide the contour into several segments and calculate the average curvature and directional characteristics of each segment; Extracting geometric shape features of the contour to describe the overall shape of the contour to form overall feature information, and statistically analyzing the number feature information of contour points, the length feature information of the contour, and the curvature feature information of the contour; The overall feature information, the number feature information of the contour points, the length feature information of the contour and the curvature feature information of the contour are combined into a feature vector to represent the contour features of the Chinese character; The feature vector is input into the trained classifier for identification to determine the font type to which it belongs; the above-mentioned font types include regular script, running script, official script, and small seal script.

[0011] Based on the first aspect, the specific steps of inputting Chinese characters into the corresponding preset Chinese character review model according to the font type to analyze the radical, component and overall structure of the Chinese characters to obtain the review results and display them include: According to the font type, Chinese characters are input into Baidu Smart Cloud Qianfan Big Model Platform to analyze the radicals, components and overall structure of the Chinese characters to obtain the evaluation results and display them.

[0012] In a second aspect, the present application provides a Chinese character writing review system, comprising: An acquisition module is used to acquire the Chinese character image to be reviewed and extract a single Chinese character image therefrom; A preprocessing module, used for preprocessing any single Chinese character image to obtain a preliminary processed image; An extraction module, used to extract the outline of Chinese characters in the preliminary processed image using an edge detection algorithm; The display module is used to determine the font type of the Chinese character based on the outline of the Chinese character, and input the Chinese character into the corresponding pre-set Chinese character comment model according to the font type to analyze the radical, component and overall structure of the Chinese character to obtain the comment result and display it.

[0013] In a second aspect, the present application provides an electronic device, including: A memory for storing one or more programs; processor; When the one or more programs are executed by the processor, the method is implemented.

[0014] Compared with the prior art, the present invention has at least the following advantages or beneficial effects: By preprocessing the extracted single Chinese character image, the image quality is improved to make it more suitable for subsequent analysis. The outline of the Chinese character is accurately outlined through edge detection, and the font type of the Chinese character is determined based on the outline of the Chinese character. According to the determined font type of the Chinese character, the Chinese character image is input into a pre-trained review model that matches the font type. The review model conducts a detailed analysis of the input Chinese character image, evaluates the accuracy of its radicals and components, and the rationality of the overall structure, thereby generating a review result, including the correctness of the Chinese character writing, the degree of beauty, and the evaluation of the structural balance, and displays this result to the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A flowchart of an embodiment of a method for reviewing Chinese character writing according to the present application; Figure 2 A flowchart of an embodiment of a method for reviewing Chinese character writing according to the present application; Figure 3 A flowchart of an embodiment of a method for reviewing Chinese character writing according to the present application; Figure 4 This is a structural diagram of an embodiment of a Chinese character writing review system of the present application; Figure 5 A structural block diagram of an embodiment of an electronic device of the present application. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0018] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0019] Example The embodiments of the present application provide a method and system for commenting on Chinese character writing, which can increase the accuracy of commenting on Chinese character writing.

[0020] See also Figure 1 In a first aspect, the present application provides a method for reviewing Chinese character writing, comprising the following steps: S1: Obtain the Chinese character image to be reviewed and extract a single Chinese character image from it; Specifically, the method for obtaining the Chinese character image to be reviewed includes: establishing a Chinese character review application, which is equivalent to establishing an APP or a small program. Adding network, camera and storage permissions in the created project file; In order to allow the APP to access protected resources on the device (such as a mobile phone), such as the camera and storage space, so that users can upload pictures, we add network, camera and storage permissions in the created project file. Among them, the network permission allows the APP to access the network for sending and receiving data. Since our APP needs to call the pre-set Chinese character review model (such as the Wenxin Yiyan API), the network permission is added. The camera permission allows the APP to access the device's camera for taking pictures, and the user can directly take pictures and upload pictures smoothly. The storage permission allows the APP to read and write external storage space, and the user can directly upload pictures stored in the mobile phone. The review interface of the Chinese character review application has an image upload icon and a photo icon; when it is detected that the image upload icon is selected, an image upload request is generated, and according to the image upload request, it jumps to the image selection interface, and checks the network permission at the same time. If it passes, the selected image is marked as the Chinese character image to be reviewed and uploaded to the application; it is equivalent to that the user can directly upload the picture for recognition, and the uploaded picture can be a pre-prepared picture. When it is detected that the photo icon is selected, a photo request is generated and the network permission is checked. If it is passed, the camera is opened to take photos, and the image taken this time is transmitted to the image display interface. After the photo is detected, it jumps to the image display interface to wait for selection and confirmation; after the selection is confirmed, it is marked as a Chinese character image to be commented on and uploaded to the application. It should be noted that in addition to directly uploading pictures for recognition, users can also take pictures and recognize them at the same time, that is, click the photo icon to take pictures. If only one picture is taken, it will be automatically recognized. If multiple pictures are taken, it will jump to the image display interface after the photo is taken to wait for selection and confirmation. Users do not need to open the gallery to re-select, which is convenient and quick.

[0021] See also Figure 2 , the specific steps of extracting a single Chinese character image include: S1-1: Use OpenCV's contour detection algorithm to find all contours in the image; specifically, OpenCV provides a series of image processing algorithms, including contour detection. The contour detection algorithm can identify the edges of Chinese characters in the image, thereby outlining the contours of the object.

[0022] S1-2: According to the size and shape of the outline, the outline representing the Chinese character is selected; specifically, after all the outlines are identified, the outline representing the Chinese character is selected according to the characteristics of the outline. Since Chinese characters usually have a certain size range and specific shape characteristics, the outlines that meet the conditions can be selected by setting the perimeter, area, length and width of the outline, and these outlines are likely to be Chinese characters.

[0023] S1-3: Segment the image into individual Chinese character images based on the filtered contours.

[0024] After filtering out the contours representing Chinese characters, the original image is segmented according to these contours to obtain each individual Chinese character image. The original image is cropped using the contour information (such as the bounding box of the contour), and each cropped area contains a single Chinese character.

[0025] S2: pre-process any single Chinese character image to obtain a preliminary processed image; See also Figure 3 Specifically, the specific steps of preprocessing any single Chinese character image to obtain a preliminary processed image include: S2-1: Process any single Chinese character image and convert it into a grayscale image; specifically, convert the color information of each pixel in the image into a grayscale value. Grayscale images contain less information than color images and are faster and easier to process.

[0026] S2-2: Convert the grayscale image to a binary image; specifically, select a threshold, set all pixels in the grayscale image whose grayscale values ​​are lower than the threshold to 0 (black), and set all pixels whose grayscale values ​​are higher than or equal to the threshold to 255 (white). In this way, the image is converted into a binary image containing only black and white colors.

[0027] S2-3: Use a filtering algorithm to remove noise from a binary image to obtain a preliminary processed image. The specific steps include: determining the type of noise in the binary image, which includes one or more of salt and pepper noise, speckle noise, Gaussian noise, small noise points, and holes; for salt and pepper noise and speckle noise, select a corresponding filter according to the size of the image to replace the pixel value with the median value of the pixel value in its neighborhood; for Gaussian noise, select a corresponding filter according to the size of the image to replace the pixel value with the average value of the pixel value in its neighborhood; for small noise points and holes, perform an erosion operation on the image to remove small noise points, and perform an expansion operation on the image to fill the holes. Erosion operation: By setting a structural element (such as a small rectangle or circle), and then traversing each pixel in the image, if all the pixels in the structural element are foreground pixels (usually white in a binary image), the pixel point is retained as a foreground pixel; otherwise, the pixel point is set to a background pixel (black). In this way, small noise points can be removed. Dilation operation: In contrast to the erosion operation, the dilation operation checks the pixels within the structural element, and if at least one of them is a foreground pixel, the pixel at the center of the structural element is set as a foreground pixel. This can fill the holes in the image.

[0028] S3: Use edge detection algorithm to extract the outline of Chinese characters in the preliminary processed image; Specifically, the function cv2.Gaussian Blur(gray, (5, 5), 0) is used for blurring to obtain a blurred image, where (5, 5) is the size of the blur kernel. Through such processing, the noise and details in the image are reduced, making edge detection more accurate; the function cv2.Canny(gray, 75, 200) is used for Canny edge detection to obtain an edge detection image, where 75 and 200 are threshold parameters; the edges of the edge detection image are represented by white, and the rest are represented by black. The cv2.findContours() function is used to find the contours in the edge detection image; according to the area and shape characteristics of the contours, the contours of the Chinese characters are filtered out, and a list of all detected contours is returned; the cv2.contourArea() function is used to calculate the area of ​​the contour, and the conditional statement is used to filter to obtain the filtered image, the filtered contour list, these contours are considered to represent Chinese characters; the cv2.drawContours() function is used to draw the filtered contours on the filtered image. The purpose of this setting is to draw the filtered contours on the original image or the filtered image to obtain an image containing the drawn Chinese character contours.

[0029] S4: Determine the font type of the Chinese character based on the outline of the Chinese character, and input the Chinese character into the corresponding pre-set Chinese character review model according to the font type to analyze the radical, component and overall structure of the Chinese character to obtain the review result and display it.

[0030] Specifically, the specific steps of judging the font type of a Chinese character based on the outline of the Chinese character include: dividing the outline into several segments, and calculating the average curvature and directional features of each segment; extracting the geometric shape features of the outline to describe the overall shape of the outline to form overall feature information, and counting the number feature information of the outline points, the length feature information of the outline, and the curvature feature information of the outline; combining the overall feature information, the number feature information of the outline points, the length feature information of the outline, and the curvature feature information of the outline into a feature vector for representing the outline features of the Chinese character; inputting the feature vector into the trained classifier for recognition to determine the font type to which it belongs; the font types include regular script, running script, official script, and small seal script. According to the font type, the Chinese characters are input into the Baidu Smart Cloud Qianfan Big Model Platform to analyze the radicals, components, and overall structure of the Chinese characters to obtain the comment results and display them. Through such a setting, the recognized Chinese characters are input into the Baidu Smart Cloud Qianfan Big Model Platform. The platform uses advanced natural language processing technology and deep learning models to conduct a deeper analysis of Chinese characters and identify the radicals, components, and overall structure of Chinese characters. Generate the review results of Chinese character structure and display them through a suitable interface or method. In the Baidu Smart Cloud Qianfan Big Model Platform website (https: / / qianfan.cloud.baidu.com / ), first create a new application in the Baidu Smart Cloud Qianfan console to obtain the AppID, APIKey and SecretKey of the corresponding AI model. These information will be used for the authentication process of API calls. Then use the API Key and Secret Key to obtain the interface access credential (access_token). In the app development code, you first need to build an API request and prepare an HTTP request, including setting the correct request header and request body. The request body contains the image data uploaded by the user and the prompt words given above, as well as any other necessary parameters. The second is to send a request. Use a network library (such as OkHttp) to build and send a request to the Wenxin Yiyan API, where the request may contain the Base64 encoded data of the image or the URL of the image, as well as the relevant prompt words. Finally, perform response processing, that is, receive and parse the response returned by the Wenxin Yiyan server. The response is usually in JSON format and contains the results of API processing. Displaying the above parsed results to the user, that is, returning the final result to the corresponding position of the app page, will involve updating the user interface of the app to display the text or other information returned by Wenxin Yiyan.

[0031] See also Figure 4In the second aspect, the present application provides a Chinese character writing review system, comprising: The acquisition module 1 is used to acquire the Chinese character image to be commented on and extract a single Chinese character image from it; the preprocessing module 2 is used to preprocess any single Chinese character image to obtain a preliminary processed image; the extraction module 3 is used to extract the outline of the Chinese character in the preliminary processed image using an edge detection algorithm; the display module 4 is used to determine the font type of the Chinese character based on the outline of the Chinese character, and input the Chinese character into the corresponding pre-set Chinese character comment model according to the font type to analyze the radical, component and overall structure of the Chinese character to obtain the comment result and display it. For the specific implementation of the comment system, please refer to the specific implementation of the comment method above, and no further details will be given here.

[0032] See also Figure 5 In a second aspect, the present application provides an electronic device, comprising: a memory 6 for storing one or more programs; a processor 5; when the one or more programs are executed by the processor 5, the above-mentioned Chinese character writing comment method is implemented. The processor 5 and the memory 6 transmit data via a communication interface 7.

[0033] It will be apparent to those skilled in the art that the present application is not limited to the details of the exemplary embodiments described above, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims be included in the present application. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for commenting on Chinese character writing, characterized in that: include: Obtain the Chinese character image to be reviewed, and extract a single Chinese character image from it; Preprocess any single Chinese character image to obtain a preliminary processed image; Use edge detection algorithm to extract the outline of Chinese characters in the preliminary processed image; The font type of the Chinese character is determined based on the outline of the Chinese character. According to the font type, the Chinese character is input into the corresponding pre-set Chinese character review model to analyze the radical, component and overall structure of the Chinese character to obtain the review result and display it.

2. A method for commenting on Chinese character writing according to claim 1, characterized in that: The step of obtaining the Chinese character image to be reviewed specifically includes: Establishing a Chinese character review application, adding network, camera and storage permissions in the created project file; the review interface of the Chinese character review application has an image upload icon and a photo taking icon; When it is detected that the image upload icon is selected, an image upload request is generated, and the image selection interface is jumped to according to the image upload request. At the same time, the network permission is checked. If it passes, the selected image is marked as the Chinese character image to be reviewed and uploaded to the application; When it is detected that the photo icon is selected, a photo request is generated and the network permission is checked. If it is passed, the camera is opened to take a photo, and the image taken this time is transmitted to the image display interface. After the photo is completed, it jumps to the image display interface to wait for selection. After the selection is confirmed, it will be marked as a Chinese character image to be commented on and uploaded to the application.

3. A method for commenting on Chinese character writing according to claim 1, characterized in that: The specific steps of extracting a single Chinese character image therefrom include: Use OpenCV's contour detection algorithm to find all contours in the image; According to the size and shape of the outline, the outline representing the Chinese character is selected; According to the filtered contours, the image is segmented into individual Chinese character images.

4. A method for commenting on Chinese character writing according to claim 1, characterized in that: The specific steps of preprocessing any single Chinese character image to obtain a preliminary processed image include: Process any single Chinese character image and convert it into a grayscale image; Convert grayscale image to binary image; A filtering algorithm is used to remove noise from the binary image to obtain a preliminary processed image.

5. A method for commenting on Chinese character writing according to claim 4, characterized in that: The specific steps of using the filtering algorithm to remove noise in the binary image include: Determine the noise type in the binary image, where the noise type includes one or more of salt and pepper noise, speckle noise, Gaussian noise, small noise points, and holes; For salt and pepper noise and speckle noise, select the corresponding filter according to the size of the image and replace the pixel value with the median of the pixel values ​​in its neighborhood for each pixel in the image; For Gaussian noise, the corresponding filter is selected according to the size of the image. For each pixel in the image, the pixel value is replaced by the average value of the pixel values ​​in its neighborhood. For small noise points and holes, the image is eroded to remove the small noise points, and the image is expanded to fill the holes.

6. A method for commenting on Chinese character writing according to claim 5, characterized in that: The specific steps of extracting the outline of Chinese characters in the preliminary processed image using an edge detection algorithm include: Use the function cv2.Gaussian Blur(gray, (5, 5), 0) to blur the image, where (5, 5) is the size of the blur kernel. Use the function cv2.Canny(gray, 75, 200) to perform Canny edge detection to obtain the edge detection image, where 75 and 200 are threshold parameters; Use cv2.findContours() function to find contours in edge detection image; Filter out the outline of Chinese characters according to the area and shape characteristics of the outline; Use cv2.contourArea() function to calculate the area of ​​the contour, and use conditional statements to filter to get the filtered image; Draw the filtered contours on the filtered image using cv2.drawContours() function.

7. A method for commenting on Chinese character writing according to claim 6, characterized in that: The specific steps of determining the font type of a Chinese character based on the outline of the Chinese character include: Divide the contour into several segments and calculate the average curvature and directional characteristics of each segment; Extracting geometric shape features of the contour to describe the overall shape of the contour to form overall feature information, and statistically analyzing the number feature information of contour points, the length feature information of the contour, and the curvature feature information of the contour; The overall feature information, the number feature information of the contour points, the length feature information of the contour and the curvature feature information of the contour are combined into a feature vector to represent the contour features of the Chinese character; The feature vector is input into the trained classifier for identification to determine the font type to which it belongs; the font types include regular script, running script, official script, and small seal script.

8. A method for commenting on Chinese character writing according to claim 1, characterized in that: The specific steps of inputting Chinese characters into the corresponding preset Chinese character review model according to the font type to analyze the radical, component and overall structure of the Chinese characters to obtain the review results and display them include: According to the font type, Chinese characters are input into Baidu Smart Cloud Qianfan Big Model Platform to analyze the radicals, components and overall structure of the Chinese characters to obtain the evaluation results and display them.

9. A Chinese character writing review system, characterized in that: include: An acquisition module is used to acquire the Chinese character image to be reviewed and extract a single Chinese character image therefrom; A preprocessing module, used for preprocessing any single Chinese character image to obtain a preliminary processed image; An extraction module, used to extract the outline of Chinese characters in the preliminary processed image using an edge detection algorithm; The display module is used to determine the font type of the Chinese character based on the outline of the Chinese character, and input the Chinese character into the corresponding pre-set Chinese character comment model according to the font type to analyze the radical, component and overall structure of the Chinese character to obtain the comment result and display it.

10. An electronic device, characterized in that: include: A memory for storing one or more programs; processor; When the one or more programs are executed by the processor, the method according to any one of claims 1 to 8 is implemented.

Images