Sensitive statement detection method and system

Abstract

The invention provides a sensitive statement detection method and system. The method comprises the following steps: receiving to-be-detected content uploaded by a user; obtaining to-be-detected text data according to the to-be-detected content; dividing the to-be-detected text data into at least one detection unit; the detection unit is a piece of text data with complete sentence meaning; for each detection unit, carrying out sensitive word retrieval in a sensitive word database, and determining whether the detection unit is a sensitive statement or not; converting each detection unit into a vector, performing similarity matching on each converted vector in a sensitive vector database, and determining whether the detection unit is a sensitive statement or not; and outputting a retrieval result of the sensitive statement. According to the method, sensitive statements on the semantic level can be processed, and the detection accuracy and efficiency are improved.

Description

Technical Field

[0001] One or more embodiments of the present invention relate to data processing technology, and more particularly, to a method and system for detecting sensitive sentences. Background Art

[0002] In the digital age, the Internet is developing rapidly, and the speed and breadth of information dissemination have reached unprecedented heights. Effectively detecting, managing, and filtering sensitive words or sentences in text information is an important part of the data security system. For example, in applications that have embedded large models, text interaction with large models is essential. Adding sensitive sentence detection steps can prevent large models from generating language for sensitive inputs or generating more sensitive language.

[0003] At present, sensitive sentences are mainly detected by manually matching keywords in text data. In the company's internal information system, a large amount of text data needs to be transmitted. If manual detection is performed every time it is transmitted, work efficiency will be greatly reduced, and the detection accuracy of large texts cannot be guaranteed. Summary of the invention

[0004] One or more embodiments of the present invention describe a method and system for detecting sensitive sentences, which can improve detection efficiency and detection accuracy.

[0005] According to a first aspect, a method for detecting sensitive sentences is provided, the method comprising:

[0006] The predetermined sensitive words are stored in a sensitive word database;

[0007] Convert predetermined sentences with sensitive semantics into vectors, and save the vectors in a sensitive vector database;

[0008] The method further includes:

[0009] Receiving content to be detected;

[0010] Obtaining text data to be detected according to the content to be detected;

[0011] Divide the text data to be detected into at least one detection unit; the detection unit is: a section of text data with complete sentence meaning;

[0012] For each detection unit, a sensitive word search is performed in the sensitive word database to determine whether the detection unit is a sensitive sentence;

[0013] Convert each detection unit into a vector, perform similarity matching on each converted vector in a sensitive vector database, and determine whether the detection unit is a sensitive sentence;

[0014] Output the search results of sensitive sentences.

[0015] The content to be detected is: a file in docx format, doc format, txt format or pdf format output by the file system of the information system;

[0016] or,

[0017] The content to be detected is: text data output by the artificial intelligence model.

[0018] When the content to be detected is a file in PDF format; the step of obtaining the text data to be detected according to the content to be detected includes:

[0019] The pdf format file is converted into text data in docx, doc or txt format, and the converted docx, doc or txt format text data is used as the text data to be detected.

[0020] The step of dividing the text data to be detected into at least one detection unit comprises:

[0021] The text data to be detected is segmented using commas, periods, question marks, exclamation marks, semicolons and ellipsis marks as segmentation marks to obtain at least one detection unit.

[0022] The step of searching for sensitive words in a sensitive word database for each detection unit includes:

[0023] Perform word segmentation on each detection unit, thereby obtaining a word segmentation set corresponding to the detection unit including at least one word;

[0024] Traverse each word in the word set and search for each word in the word set in the sensitive word database;

[0025] If a word in a word set can be retrieved in the sensitive word database, then the detection unit corresponding to the word set is determined as a sensitive sentence.

[0026] The step of searching each converted vector in the sensitive vector database includes:

[0027] For the vector corresponding to each detection unit, similarity matching is performed between each converted vector and each vector stored in the sensitive vector database according to a pre-set similarity threshold;

[0028] If the similarity between a converted vector and a vector stored in the sensitive vector database is greater than a similarity threshold, the detection unit corresponding to the converted vector is determined as a sensitive sentence.

[0029] The output sensitive sentence search results include:

[0030] Record each detection unit determined to be a sensitive sentence in a table item in an Excel table;

[0031] Output the Excel table.

[0032] The output sensitive sentence search results include:

[0033] If the content to be detected is a file, the marked file is further outputted, in which each detection unit determined to be a sensitive sentence is marked.

[0034] According to a second aspect, a sensitive sentence detection system is provided, the system comprising:

[0035] A sensitive word database, used to store pre-determined sensitive words;

[0036] A sensitive vector database, used to store vectors converted from pre-determined sentences with sensitive semantics;

[0037] The to-be-detected data acquisition module is configured to receive the to-be-detected content; and obtain the to-be-detected text data according to the to-be-detected content;

[0038] A division module is configured to divide the text data to be detected into at least one detection unit; the detection unit is: a section of text data with complete sentence meaning;

[0039] A first sensitive sentence processing module is configured to perform a sensitive word search in a sensitive word database for each detection unit to determine whether the detection unit is a sensitive sentence;

[0040] A second sensitive sentence processing module is configured to convert each detection unit into a vector, perform similarity matching on each converted vector in a sensitive vector database, and determine whether the detection unit is a sensitive sentence;

[0041] The output module is configured to output the retrieval results of the sensitive sentences according to the sensitive sentence determination results of the first sensitive sentence processing module and the second sensitive sentence processing module.

[0042] According to a third aspect, a computing device is provided, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method described in any embodiment of the present invention is implemented.

[0043] The sensitive sentence detection method and system provided by various embodiments of the present invention have at least the following beneficial effects:

[0044] 1. In the embodiment of the present invention, the system can automatically detect sensitive sentences without human intervention, thereby greatly improving work efficiency and ensuring the detection accuracy of large texts.

[0045] 2. The present invention can process sensitive sentences at the semantic level and improve the accuracy and efficiency of detection. In an embodiment of the present invention, the text data to be detected is divided into various detection units, and each detection unit is a section of text data with complete sentence meaning, that is, each detection unit can correspond to a sentence, so that the vector corresponding to the detection unit is matched with the vector of the sentence with sensitive semantics pre-stored in the sensitive vector database for similarity. If the match is successful, it means that the matched detection unit has sensitive semantics. Therefore, even if the detection unit does not include sensitive words, the detection unit with sensitive semantics can be accurately determined, that is, the sentence with sensitive semantics can be detected. Therefore, the embodiment of the present invention further improves the accuracy of sensitive detection.

[0046] 3. In the embodiment of the present invention, the content to be detected can be: a file in docx format, doc format, txt format or pdf format output by the file system of the information system. In this way, sensitive sentences can be directly detected for files output by various information systems, such as the company's information query system, thus expanding the detection scope.

[0047] 4. In the embodiment of the present invention, the content to be detected can be text data output by the artificial intelligence model. In this way, it is possible to perform sensitive detection on the text data output by various artificial intelligence models. For example, if the question-answering assistant intelligent model outputs the answer in the form of text data to a question raised by a user, the embodiment of the present invention can perform sensitive detection on the answer.

[0048] 5. After using the method of the embodiment of the present invention to detect sensitive sentences, each information system can improve data quality and return sensitive files when receiving files uploaded by users; filter sensitive questions and answers during the interaction with the large model to avoid disputes, which helps to improve the security of the use of artificial intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 It is a flow chart of a method for detecting sensitive sentences in one embodiment of the present invention.

[0051] Figure 2 It is a structural diagram of a sensitive sentence detection system in one embodiment of the present invention.

[0052] Figure 3 It is an application schematic diagram of a sensitive sentence detection system in one embodiment of the present invention. DETAILED DESCRIPTION

[0053] The solution provided by the present invention is described below in conjunction with the accompanying drawings.

[0054] First of all, it should be noted that the terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates otherwise.

[0055] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0056] Figure 1 FIG. 1 is a flow chart of a method for detecting sensitive sentences in one embodiment of the present invention. Figure 1 , the method comprising:

[0057] Step 101: Save the pre-determined sensitive words in a sensitive word database;

[0058] Step 103: converting the predetermined sentence with sensitive semantics into a vector, and storing the vector in a sensitive vector database;

[0059] Step 105: receiving content to be detected;

[0060] Step 107: obtaining text data to be detected according to the content to be detected;

[0061] Step 109: Divide the text data to be detected into at least one detection unit; the detection unit is: a section of text data with complete sentence meaning;

[0062] Step 111: for each detection unit, perform a sensitive word search in a sensitive word database to determine whether the detection unit is a sensitive sentence;

[0063] Step 113: convert each detection unit into a vector, perform similarity matching on each converted vector in a sensitive vector database, and determine whether the detection unit is a sensitive sentence;

[0064] Step 115: Output the search results of sensitive sentences.

[0065] according to Figure 1 It can be seen from the sensitive sentence detection method of the present invention that in the embodiment of the present invention, the system can automatically complete the detection of sensitive sentences without human intervention, thereby greatly improving work efficiency and ensuring the detection accuracy of large texts.

[0066] In the prior art, usually only the sensitive words included in a sentence are detected to determine whether a sentence is a sensitive sentence. For example, if a sentence includes sensitive words such as "privacy", "ID number" or "discrimination", then the sentence can be determined to be a sensitive sentence. However, in many scenarios, even if sensitive words do not appear in a sentence, the sentence may still have sensitive semantics, such as "Please tell me the number of a card that can prove your information." If the keyword / sensitive word detection of the prior art is used, it is impossible to accurately determine whether the sentence that does not include sensitive words is a sensitive sentence. In other words, the prior art cannot achieve sensitivity detection at the semantic level. According to Figure 1 The method shown, in an embodiment of the present invention, divides the text data to be detected into various detection units, each detection unit is a section of text data with complete sentence meaning, that is, each detection unit can correspond to a sentence, that is, can correspond to a semantics, so that the vector corresponding to the detection unit is matched with the vector of the sentence with sensitive semantics pre-stored in the sensitive vector database for similarity. If the match is successful, it means that the matched detection unit has sensitive semantics. Therefore, even if the detection unit does not include sensitive words, the detection unit with sensitive semantics can be accurately determined, that is, the sentence with sensitive semantics can be detected. Therefore, the embodiment of the present invention further improves the accuracy of sensitive detection.

[0067] Combine the following Figure 1 Each step shown illustrates the process of the sensitive sentence detection method in the embodiment of the present invention.

[0068] First, step 101 is executed: pre-determined sensitive words are saved in a sensitive word database.

[0069] Based on experience and compliance requirements, sensitive words are saved in the sensitive word database in advance. For example, sensitive words include: "privacy", "ID number" and "discrimination".

[0070] In the embodiment of the present invention, es may be used to maintain the sensitive word database.

[0071] Execute step 103: convert the predetermined sentence with sensitive semantics into a vector, and save the vector in a sensitive vector database.

[0072] In an embodiment of the present invention, faiss may be used to maintain a sensitive vector database, wherein the vector data stored in the sensitive vector database is generated by a BGE text vectorization model.

[0073] Step 105: Receive the content to be detected.

[0074] In one embodiment of the present invention, the content to be detected can be received from multiple channels, that is, sensitive detection can be performed on the content in multiple channels, which expands the application of the service and improves the user satisfaction.

[0075] For example, in one embodiment of the present invention, the content to be detected is: the content to be detected directly uploaded by the user. The content to be detected uploaded by the user can be a file in various formats, or text data such as a paragraph of text. In this way, sensitive detection of the content uploaded by the user can be achieved.

[0076] For example, see Figure 3 In one embodiment of the present invention, the content to be detected can be: a file in docx format, doc format, txt format or pdf format output by the file system of the information system. In this way, sensitive sentences can be directly detected for files output by various information systems, such as the company's information query system, thus expanding the detection scope.

[0077] For example, see Figure 3 In one embodiment of the present invention, the content to be detected can be text data output by an artificial intelligence model. In this way, sensitive detection can be performed on text data output by various artificial intelligence models. For example, if a question-answering assistant intelligent model outputs an answer in the form of text data to a question raised by a user, the present invention can perform sensitive detection on the answer.

[0078] Next, step 107 is executed: obtaining text data to be detected according to the content to be detected.

[0079] The received content to be detected may be in various formats, so the formats need to be unified into text format.

[0080] When the content to be detected is a file in PDF format; in this step 107, the text data to be detected is obtained according to the content to be detected, including:

[0081] The pdf format file is converted into text data in docx, doc or txt format, and the converted docx, doc or txt format text data is used as the text data to be detected.

[0082] Next, step 109 is performed: dividing the text data to be detected into at least one detection unit; the detection unit is: a section of text data with complete sentence meaning.

[0083] In this step 109, the text data to be detected is divided into at least one detection unit, including:

[0084] The text data to be detected is segmented using commas, periods, question marks, exclamation marks, semicolons, and ellipsis marks as segmentation marks to obtain at least one detection unit. In this way, the obtained detection unit can be a piece of text data with complete sentence meaning, and subsequently, sensitive detection at the semantic level can be achieved.

[0085] Next, step 111 is executed: for each detection unit, a sensitive word search is performed in the sensitive word database to determine whether the detection unit is a sensitive sentence.

[0086] In step 111, the process of searching for sensitive words in the sensitive word database for each detection unit may include:

[0087] Step 1111: performing word segmentation on each detection unit, thereby obtaining a word segmentation set corresponding to the detection unit including at least one word;

[0088] Step 1112: traverse each word in the word set, and search each word in the word set in the sensitive word database;

[0089] Step 1113: If a word in the word set can be retrieved in the sensitive word database, then the detection unit corresponding to the word set is determined as a sensitive sentence.

[0090] Next, step 113 is performed: each detection unit is converted into a vector, and each converted vector is matched for similarity in a sensitive vector database to determine whether the detection unit is a sensitive sentence.

[0091] In one embodiment of the present invention, searching each converted vector in the sensitive vector database in step 113 includes:

[0092] Step 1131: for each vector corresponding to each detection unit, similarity matching is performed between each converted vector and each vector stored in the sensitive vector database according to a preset similarity threshold;

[0093] Step 1133: If the similarity between a converted vector and a vector stored in the sensitive vector database is greater than a similarity threshold, the detection unit corresponding to the converted vector is determined as a sensitive sentence.

[0094] It can be seen that in the embodiment of the present invention, in the keyword search process, the detection unit is first segmented, and the jieba segmentation tool can be used. After the segmentation, the segmentation set corresponding to the detection unit is obtained, and the segmentation set is traversed to search in the sensitive keyword library. In the similarity matching process, the detection unit is first vectorized using the BGE model, and the obtained vector data is similarly matched in the sensitive vector database. It is necessary to set the similarity threshold in advance. If it is higher than the threshold, it is regarded as a sensitive sentence.

[0095] Next, execute step 115: output the search results of sensitive sentences.

[0096] In one embodiment of the present invention, outputting the search results of sensitive sentences includes:

[0097] Record each detection unit determined to be a sensitive sentence in a table item in an Excel table;

[0098] Output the Excel table.

[0099] In one embodiment of the present invention, in step 115, outputting the search results of sensitive sentences may further include:

[0100] If the content to be detected uploaded by the user is a file, the marked file is further outputted, in which each detection unit determined to be a sensitive sentence is marked, for example, the text detected as a sensitive sentence is marked in red.

[0101] It can be seen that after using the method of the embodiment of the present invention to detect sensitive sentences, each information system can improve data quality and return sensitive files when receiving files uploaded by users; filter sensitive questions and answers during the interaction with the large model to avoid disputes, which helps to improve the security of the use of artificial intelligence.

[0102] An embodiment of the present invention provides a sensitive sentence detection system, see Figure 2 The system comprises:

[0103] Sensitive word database 201, used to store pre-determined sensitive words;

[0104] A sensitive vector database 202, used to store vectors converted from predetermined sentences with sensitive semantics;

[0105] The to-be-detected data acquisition module 203 is configured to receive the to-be-detected content; and obtain the to-be-detected text data according to the to-be-detected content;

[0106] The division module 204 is configured to divide the text data to be detected into at least one detection unit; the detection unit is: a section of text data with complete sentence meaning;

[0107] The first sensitive sentence processing module 205 is configured to search for sensitive words in the sensitive word database for each detection unit to determine whether the detection unit is a sensitive sentence;

[0108] The second sensitive sentence processing module 206 is configured to convert each detection unit into a vector, perform similarity matching on each converted vector in the sensitive vector database, and determine whether the detection unit is a sensitive sentence;

[0109] The output module 207 is configured to output the retrieval results of the sensitive sentences according to the sensitive sentence determination results of the first sensitive sentence processing module and the second sensitive sentence processing module.

[0110] In one embodiment of the system of the present invention, the content to be detected is: a file in docx format, doc format, txt format or pdf format output by a file system of an information system;

[0111] or,

[0112] In one embodiment of the system of the present invention, the content to be detected is: text data output by the artificial intelligence model.

[0113] In one embodiment of the system of the present invention, when the content to be detected is a file in PDF format; the to-be-detected data acquisition module 203 is configured to execute:

[0114] The pdf format file is converted into text data in docx, doc or txt format, and the converted docx, doc or txt format text data is used as the text data to be detected.

[0115] In one embodiment of the system of the present invention, the partitioning module 204 is configured to perform:

[0116] The text data to be detected is segmented using commas, periods, question marks, exclamation marks, semicolons and ellipsis marks as segmentation marks to obtain at least one detection unit.

[0117] In one embodiment of the system of the present invention, the first sensitive sentence processing module 205 is configured to execute:

[0118] Perform word segmentation on each detection unit, thereby obtaining a word segmentation set corresponding to the detection unit including at least one word;

[0119] Traverse each word in the word set and search for each word in the word set in the sensitive word database;

[0120] If a word in a word set can be retrieved in the sensitive word database, then the detection unit corresponding to the word set is determined as a sensitive sentence.

[0121] In one embodiment of the system of the present invention, the second sensitive sentence processing module 206 is configured to execute:

[0122] For the vector corresponding to each detection unit, similarity matching is performed between each converted vector and each vector stored in the sensitive vector database according to a pre-set similarity threshold;

[0123] If the similarity between a converted vector and a vector stored in the sensitive vector database is greater than a similarity threshold, the detection unit corresponding to the converted vector is determined as a sensitive sentence.

[0124] In one embodiment of the system of the present invention, the output module 207 is configured to perform:

[0125] Record each detection unit determined to be a sensitive sentence in a table item in an Excel table;

[0126] Output the Excel table.

[0127] In one embodiment of the system of the present invention, the output module 207 is configured to execute: if the content to be detected is a file, further output the marked file, in which each detection unit determined to be a sensitive sentence is marked.

[0128] It should be noted that the above modules are usually implemented on the server side, and can be set on independent servers respectively, or some or all of the modules can be set on the same server. The server can be a single server or a server cluster composed of multiple servers. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system. The above devices can also be implemented in a computer terminal with strong computing power.

[0129] An embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed in a computer, the computer is caused to execute a method in any one of the embodiments in the specification.

[0130] An embodiment of the present invention provides a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method in any embodiment of the specification is implemented.

[0131] It is to be understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on the device of the embodiment of the present invention. In other embodiments of the specification, the above-mentioned device may include more or fewer components than those shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0132] The various embodiments of the present invention are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0133] Those skilled in the art should be aware that in one or more of the above examples, the functions described in the present invention can be implemented by hardware, software, widgets, or any combination thereof. When implemented by software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0134] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for detecting sensitive sentences, characterized in that: The method includes: The predetermined sensitive words are stored in a sensitive word database; Convert predetermined sentences with sensitive semantics into vectors, and save the vectors in a sensitive vector database; The method further includes: Receiving content to be detected; Obtaining text data to be detected according to the content to be detected; Divide the text data to be detected into at least one detection unit; the detection unit is: a section of text data with complete sentence meaning; For each detection unit, a sensitive word search is performed in the sensitive word database to determine whether the detection unit is a sensitive sentence; Convert each detection unit into a vector, perform similarity matching on each converted vector in a sensitive vector database, and determine whether the detection unit is a sensitive sentence; Output the search results of sensitive sentences.

2. The method according to claim 1, characterized in that: The content to be detected is: a file in docx format, doc format, txt format or pdf format output by the file system of the information system; or, The content to be detected is: text data output by the artificial intelligence model.

3. The method according to claim 2, characterized in that When the content to be detected is a file in PDF format; the step of obtaining the text data to be detected according to the content to be detected includes: The pdf format file is converted into text data in docx, doc or txt format, and the converted docx, doc or txt format text data is used as the text data to be detected.

4. The method according to claim 1, characterized in that The step of dividing the text data to be detected into at least one detection unit comprises: The text data to be detected is segmented using commas, periods, question marks, exclamation marks, semicolons and ellipsis marks as segmentation marks to obtain at least one detection unit.

5. The method according to claim 1, characterized in that The step of searching for sensitive words in a sensitive word database for each detection unit includes: Perform word segmentation on each detection unit, thereby obtaining a word segmentation set corresponding to the detection unit including at least one word; Traverse each word in the word set and search for each word in the word set in the sensitive word database; If a word in a word set can be retrieved in the sensitive word database, then the detection unit corresponding to the word set is determined as a sensitive sentence.

6. The method according to claim 1, characterized in that The step of searching each converted vector in the sensitive vector database includes: For the vector corresponding to each detection unit, each converted vector is matched with each vector stored in the sensitive vector database for similarity; If the similarity between a converted vector and a vector stored in the sensitive vector database is greater than a similarity threshold, the detection unit corresponding to the converted vector is determined as a sensitive sentence.

7. The method according to any one of claims 1 to 6, characterized in that: The output sensitive sentence search results include: Record each detection unit determined to be a sensitive sentence in a table item in an Excel table; Output the Excel table.

8. The method according to claim 7, characterized in that The output sensitive sentence search results include: If the content to be detected is a file, the marked file is further outputted, in which each detection unit determined to be a sensitive sentence is marked.

9. A sensitive sentence detection system, characterized in that: The system includes: A sensitive word database, used to store pre-determined sensitive words; A sensitive vector database, used to store vectors converted from pre-determined sentences with sensitive semantics; The to-be-detected data acquisition module is configured to receive the to-be-detected content; and obtain the to-be-detected text data according to the to-be-detected content; A division module is configured to divide the text data to be detected into at least one detection unit; the detection unit is: a section of text data with complete sentence meaning; A first sensitive sentence processing module is configured to perform a sensitive word search in a sensitive word database for each detection unit to determine whether the detection unit is a sensitive sentence; A second sensitive sentence processing module is configured to convert each detection unit into a vector, perform similarity matching on each converted vector in a sensitive vector database, and determine whether the detection unit is a sensitive sentence; The output module is configured to output the retrieval results of the sensitive sentences according to the sensitive sentence determination results of the first sensitive sentence processing module and the second sensitive sentence processing module.

10. A computing device, comprising a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method according to any one of claims 1 to 8 is implemented.

Images