Front-end page anomaly detection method, device, electronic device and medium

Abstract

The embodiment of the present invention relates to the field of front-end testing technology, and provides a front-end page anomaly detection method, device, electronic device and medium, the method comprising: obtaining a video of a user to be detected accessing a front-end page after privacy protection processing; extracting elements from each frame of the video of the user accessing the front-end page to obtain visual elements of each frame; classifying each frame based on the visual elements of each frame to obtain multiple frame pages corresponding to multiple page categories; performing anomaly detection on multiple frame pages corresponding to each page category based on the page element features of each page category. Thus, under the premise of protecting user privacy, based on the data of the front-end display screen recording, unsupervised and supervised machine learning, business rules, etc. are used to implement front-end anomaly detection based on a combination strategy, so as to identify abnormal pages and abnormal jumps in the entire link under the premise of protecting user privacy, thereby automatically mining the front-end display anomaly.

Description

Technical Field

[0001] The present invention relates to the field of front-end testing technology, and in particular to a front-end page anomaly detection method, device, electronic equipment and medium. Background Art

[0002] Due to some regulatory requirements, some companies need to keep records of user operations. Under the premise of hiding user sensitive information, such as ID card, mobile phone number, name and other details, the record data can show the detailed operation records of users in the software or web pages, as well as the pages that users can see. If it is possible to mine based on these screen recording-like data, abnormal pages or abnormal jumps in the entire link can be identified, thereby automatically mining front-end display anomalies and improving the automation and coverage of front-end testing.

[0003] At present, the object detection algorithms based on video are mainly related to computer vision. Their focus is often on the detection of real videos, and their use in the field of front-end display anomaly mining is relatively limited. The accuracy and recall rate of some component recognition algorithms related to HTML pages cannot be guaranteed. The current methods that rely on manual anomaly detection are inefficient, have low recognition accuracy, and have a small coverage. Therefore, how to identify page elements and page anomalies while protecting user privacy has become an urgent problem to be solved. Summary of the invention

[0004] The present invention provides a front-end page anomaly detection method, device, electronic device and medium, which are used to solve the defects of low efficiency, low recognition accuracy and small coverage of the method relying on manual anomaly detection in the prior art, and realize the identification of abnormal pages and abnormal jumps in the entire link under the premise of protecting user privacy, thereby automatically mining the front-end display anomaly.

[0005] The present invention provides a front-end page anomaly detection method, comprising:

[0006] Obtain the video of the user to be detected accessing the front-end page after privacy protection processing;

[0007] Extracting elements from each frame of the front-end page in the video of the user visiting the front-end page to obtain visual elements of each frame of the page;

[0008] Classifying each frame page based on the visual elements of each frame page to obtain a plurality of frame pages corresponding to a plurality of page categories;

[0009] Based on the page element features of each page category, anomaly detection is performed on multiple frame pages corresponding to each page category.

[0010] In a possible implementation, the method further includes:

[0011] Obtain front-end page data and user access behavior data;

[0012] Performing privacy shielding processing on the front-end page data to obtain processed front-end page data;

[0013] Restoring the processed front-end page data into an HTML page sequence based on the user access behavior data;

[0014] The HTML page sequence is rendered to obtain a video of the user accessing the page to be detected.

[0015] In a possible implementation, the method further includes:

[0016] The user privacy information in the front-end page data is shielded by business rules, and the front-end page containing the privacy information is coded to obtain the processed front-end page data.

[0017] In a possible implementation, the method further includes:

[0018] Decomposing each frame of the page in the video of the page visited by the user into elements;

[0019] After the elements are disassembled, each frame page is constructed into a first frame page containing a specific element and a second frame page not containing the specific element;

[0020] Rendering the first frame page and the second frame page into the browser, and obtaining corresponding rendered images respectively;

[0021] Determine the difference between each frame page based on the rendered images corresponding to the first frame page and the second frame page respectively;

[0022] The visual element is determined based on the difference of each frame page.

[0023] In a possible implementation, the method further includes:

[0024] Based on preset page key element features and page jump features, identifying page key elements in the visual elements;

[0025] Each frame page is classified based on the page key elements to obtain a plurality of frame pages corresponding to a plurality of page categories.

[0026] In a possible implementation, the method further includes:

[0027] Based on the page categories, determining a plurality of target key elements contained in each page category;

[0028] Based on the multiple target key elements contained in each page category, anomaly detection is performed on the multiple frame pages corresponding to each page category, wherein the anomaly detection includes missing, occlusion, and position error of the target key element.

[0029] In a possible implementation, the method further includes:

[0030] The Markov chain model is used to calculate the jump probability of each category page to other category pages;

[0031] The HTML page sequences whose jump probability is lower than a preset threshold are recorded.

[0032] The present invention also provides a front-end page anomaly detection device, comprising the following modules:

[0033] An acquisition module is used to acquire the video of the user accessing the front-end page to be detected after the privacy protection processing;

[0034] An element extraction module is used to extract elements from each frame of the front-end page video of the user visiting the front-end page to obtain visual elements of each frame of the page;

[0035] A page classification module, used for classifying each frame page based on the visual elements of each frame page, to obtain a plurality of frame pages corresponding to a plurality of page categories;

[0036] The anomaly detection module is used to perform anomaly detection on multiple frame pages corresponding to each page category based on the page element features of each page category.

[0037] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, a front-end page anomaly detection method as described above is implemented.

[0038] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the front-end page anomaly detection method as described in any one of the above is implemented.

[0039] The present invention also provides a computer program product, including a computer program, wherein when the computer program is executed by a processor, the front-end page anomaly detection method as described above is implemented.

[0040] The front-end page anomaly detection method, device, electronic device and medium provided by the present invention obtain the video of the user to be detected accessing the front-end page after the privacy protection processing; extract the elements of each frame page in the video of the user accessing the front-end page to obtain the visual elements of each frame page; classify each frame page based on the visual elements of each frame page to obtain multiple frame pages corresponding to multiple page categories; perform anomaly detection on multiple frame pages corresponding to each page category based on the page element characteristics of each page category. Compared with the defects of low efficiency, low recognition accuracy and small coverage of the method relying on manual anomaly detection in the prior art, this scheme, based on the data of the front-end display screen recording, uses unsupervised and supervised machine learning, business rules and other combined strategies to realize front-end anomaly detection, and realizes the identification of abnormal pages and abnormal jumps in the entire link under the premise of protecting user privacy, thereby automatically mining the front-end display anomaly. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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.

[0042] Figure 1 This is one of the flow charts of the front-end page anomaly detection method provided by the present invention.

[0043] Figure 2 This is the second flow chart of the front-end page anomaly detection method provided by the present invention.

[0044] Figure 3 This is the third flow chart of the front-end page anomaly detection method provided by the present invention.

[0045] Figure 4 It is a structural schematic diagram of the front-end page anomaly detection device provided by the present invention.

[0046] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] To facilitate understanding of the embodiments of the present invention, specific embodiments will be further explained below in conjunction with the accompanying drawings. The embodiments do not constitute a limitation on the embodiments of the present invention.

[0049] Figure 1 This is one of the flow charts of the front-end page anomaly detection method provided by the present invention. Figure 1 As shown, the method specifically includes:

[0050] S11, obtaining the video of the user to be detected accessing the front-end page after the privacy protection processing.

[0051] Embodiments of the present invention are combined Figure 3 As shown in detail, the embodiment of the present invention is applicable to the front-end page testing scenario. First, the front-end Hypertext Markup Language (HTML) page and the user's page access behavior data are obtained. At the source of the data, that is, before the user uploads the HTML content, a series of business rules are used to shield or process the user's privacy information. These private information include but are not limited to: user mobile phone number. The user's mobile phone number is sensitive information and needs to be hidden or replaced. ID number, ID number is an important identifier of personal identity and also needs to be protected. Name, the user's real name is also one of the private information. Keyboard input, the user's input on the page, especially the input containing sensitive information, also needs to be processed to avoid leakage. The way to process these private information may include direct deletion, replacement with placeholders, or coding in the private HTML page part (that is, blurred display so that the specific content cannot be recognized).

[0052] Furthermore, the data sent back by the front end (which has been processed for privacy protection) is input into the data warehouse. Then, through offline calculation, the data is restored into a sequence of user-operated pages and a corresponding sequence of HTML pages is generated. Next, the system renders these HTML page sequences to obtain user pages after privacy protection processing. Finally, these pages are converted into video form, that is, a video showing the user operation process (but with privacy information removed) is generated.

[0053] After the above processing, even if it is necessary to record and replay the user's operation process (for example, for customer service analysis, user behavior research, etc.), it can ensure that the user's privacy will not be leaked.

[0054] S12, extracting elements from each frame of the page in the video of the user visiting the front-end page to obtain visual elements of each frame of the page.

[0055] Each frame of the page in the video of the user visiting the page refers to an HTML page. Elements are extracted from each frame of the page. These elements may include text, pictures, buttons, links, etc. By disassembling and extracting the HTML page elements, the difference between the frames after adding or deleting a single element is obtained. Combined with the HTML path and label of the element, unsupervised methods are used for clustering. Combined with manual clustering, merging or adjustment, the clustering of visible elements on the page and their visual features are obtained. This clustering is used to obtain the visual elements of the page.

[0056] S13: Classify each frame page based on the visual elements of each frame page to obtain a plurality of frame pages corresponding to a plurality of page categories.

[0057] Based on the automatic page element extraction method, the key elements of the page are identified, and the page is classified in combination with business logic, page jump characteristics, and page links. The page jump characteristics can be the type of the previous page and the next page in the entire sequence of pages browsed by the user.

[0058] S14: performing anomaly detection on multiple frame pages corresponding to each page category based on page element features of each page category.

[0059] Based on the page’s classification and extracted page elements, identify anomalies such as the presence, occlusion, and position errors of key elements.

[0060] For example, after clustering, we can get the key elements of a page that a certain category page should include, but a specific page may have missing key elements; occlusion means that only part of a key element is displayed; position error means that a key element is in the wrong position. These situations can all be defined as page anomalies.

[0061] The front-end page anomaly detection method provided by the present invention obtains a video of a user to be detected accessing a front-end page after privacy protection processing; extracts elements from each frame of the video of the user accessing the front-end page to obtain visual elements of each frame; classifies each frame based on the visual elements of each frame to obtain multiple frame pages corresponding to multiple page categories; and performs anomaly detection on multiple frame pages corresponding to each page category based on the page element features of each page category. Compared with the defects of low efficiency, low recognition accuracy, and small coverage of the method relying on manual anomaly detection in the prior art, the present method, based on the data of the front-end display screen recording, uses unsupervised and supervised machine learning, business rules, etc. to realize front-end anomaly detection based on a combination strategy under the premise of protecting user privacy, and realizes the identification of abnormal pages and abnormal jumps in the entire link under the premise of protecting user privacy, thereby automatically mining the front-end display anomaly.

[0062] Figure 2This is the second flow chart of the front-end page anomaly detection method provided by the present invention. Figure 2 As shown, the method specifically includes:

[0063] S21. Obtain front-end page data and user access behavior data.

[0064] Embodiments of the present invention are combined Figure 3 As shown in detail, the embodiment of the present invention is applicable to the front-end page testing scenario, and first obtains the front-end Hypertext Markup Language (HTML) page and the user's page access behavior data. The user's page access behavior data refers to the various behavior data generated when the user accesses a web page or application, including but not limited to click behavior: the user's click position on the page, the number of clicks, etc. Scrolling behavior: the depth and speed of the user's scrolling page. Search behavior: the search keywords used by the user, the click status of the search results, etc. Form filling behavior: the user's filling status in the form, such as the filling time, the filling content, etc. Navigation behavior: the user's jump path between pages, return status, etc.

[0065] S22: Perform privacy shielding processing on the front-end page data to obtain processed front-end page data.

[0066] At the source of the data, that is, before the user uploads the HTML content, a series of business rules are used to shield or process the user's private information. This private information includes but is not limited to: the user's mobile phone number, which is sensitive information and needs to be hidden or replaced. The ID number, which is an important identifier of personal identity and also needs to be protected. The name, the user's real name is also one of the private information. Keyboard input, the user's input on the page, especially the input containing sensitive information, also needs to be processed to avoid leakage. The way to process this private information may include direct deletion, replacement with placeholders, or coding in the private HTML page part (that is, blurred display so that the specific content cannot be identified).

[0067] S23. Restoring the processed front-end page data into an HTML page sequence based on the user access behavior data.

[0068] Furthermore, the data sent back by the front end (which has been processed for privacy protection) is input into the data warehouse. Then, through offline calculation, this data is restored into the page sequence operated by the user and the corresponding HTML page sequence is generated.

[0069] S24, rendering the HTML page sequence to obtain a video of the user accessing the page to be detected.

[0070] Furthermore, the system will render the above HTML page sequence to obtain the user access page after privacy protection processing. Finally, these pages will be converted into video form, that is, a video showing the user operation process (but the privacy information has been removed) is generated.

[0071] After the above processing, even if it is necessary to record and replay the user's operation process (for example, for customer service analysis, user behavior research, etc.), it can ensure that the user's privacy will not be leaked.

[0072] S25, decomposing each frame of the page in the video of the page visited by the user into elements.

[0073] First, the system will disassemble the HTML page. This step involves analyzing the recorded video of the user visiting the page frame by frame and breaking each frame of the page into its constituent elements. These elements may be basic components of HTML pages such as text, images, buttons, links, etc.

[0074] S26: construct each frame page after the elements are disassembled into a first frame page containing a specific element and a second frame page not containing the specific element.

[0075] For each frame page after disassembly, two versions are constructed: one is the first frame page containing a specific element, and the other is the second frame page without the specific element. The purpose of this is to compare the differences (diff) between the two versions of the page later, especially to focus on the changes in the frame page after the addition or deletion of a specific element, so as to identify the impact of the specific element on the page.

[0076] In order to locate and process elements more accurately, the system records the HTML path (i.e. the position of the element in the HTML structure) and tags (such as , This information helps the system understand the role and importance of the element on the page.

[0077] S27: Render the first frame page and the second frame page into the browser, and obtain corresponding rendered images respectively.

[0078] The constructed first frame page and second frame page are rendered in the browser respectively, and the corresponding rendered images are captured or generated. These images will be used for subsequent image difference analysis.

[0079] S28: Determine the difference between each frame page based on the rendered images corresponding to the first frame page and the second frame page respectively.

[0080] Based on the rendered images of the first frame page and the second frame page, image difference analysis technology (such as pixel level comparison, feature extraction, etc.) is used to determine the difference caused by the presence or absence of specific elements in each frame page.

[0081] S29: Determine the visual element based on the difference between each frame page.

[0082] By analyzing the differences in each frame of the page, you can identify visual elements that have a significant impact on the appearance of the page. These elements may be the focus of user attention or an important part of the page layout.

[0083] S210: Based on preset page key element features and page jump features, identify page key elements in the visual elements.

[0084] Using the preset page key element features and page jump features, the key elements of the page are screened out from the identified visual elements. These key elements may be navigation links, important information blocks, interactive buttons, etc., which are essential for users to understand and use the page.

[0085] S211 , classifying each frame page based on the page key elements to obtain a plurality of frame pages corresponding to a plurality of page categories.

[0086] Finally, based on the presence and characteristics of the key elements of the page, the system will use unsupervised learning methods (such as K-means, DBSCAN, etc.) to classify each frame page. This process may involve machine learning algorithms or rule matching to divide pages into different categories (such as login pages, product details pages, shopping cart pages, etc.). Each category corresponds to a group of frame pages with similar key elements. Since unsupervised methods may not fully and accurately reflect the actual clustering of pages, the system also needs to allow manual merging or adjustment of clustering results.

[0087] After the above steps, the system will eventually obtain page categories and clusters of visible elements on the page and their corresponding visual features. These features may include the shape, color, size, position, etc. of the elements.

[0088] S212: Based on the page categories, determine a plurality of target key elements included in each page category.

[0089] Based on the page category, you can determine the multiple target key elements contained in each page category. For example, the target key elements of the product details page include navigation links, important information blocks, interactive buttons, etc., which are crucial for users to understand and use the page.

[0090] S213: Perform anomaly detection on multiple frame pages corresponding to each page category based on the multiple target key elements included in each page category.

[0091] In the embodiment of the present invention, anomaly detection includes missing, occluded, and position errors of target key elements. Based on multiple target key elements contained in each page category, anomaly detection is performed on multiple frame pages corresponding to each page category to identify anomalies such as the presence, occlusion, and position errors of target key elements in the frame pages.

[0092] For example, after clustering, we can get the key elements of a page that a certain category page should include, but a specific page may have missing key elements; occlusion means that only part of a key element is displayed; position error means that a key element is in the wrong position. These situations can all be defined as page anomalies.

[0093] The front-end page anomaly detection method provided by the present invention obtains a video of a user to be detected accessing a front-end page after privacy protection processing; extracts elements from each frame of the video of the user accessing the front-end page to obtain visual elements of each frame; classifies each frame based on the visual elements of each frame to obtain multiple frame pages corresponding to multiple page categories; and performs anomaly detection on multiple frame pages corresponding to each page category based on the page element features of each page category. This method, based on the data of the front-end display screen recording, uses unsupervised and supervised machine learning, business rules, etc. to realize front-end anomaly detection based on a combination strategy under the premise of protecting user privacy, and realizes the identification of abnormal pages and abnormal jumps in the entire link under the premise of protecting user privacy, thereby automatically mining the front-end display anomaly.

[0094] The front-end page anomaly detection device provided by the present invention is described below. The front-end page anomaly detection device described below and the front-end page anomaly detection method described above can be referenced to each other.

[0095] Figure 4 It is a structural diagram of the front-end page anomaly detection device provided by the present invention, which specifically includes:

[0096] The acquisition module 401 is used to acquire the video of the user accessing the front-end page to be detected after the privacy protection process. For detailed description, please refer to the relevant description corresponding to the above method embodiment, which will not be repeated here.

[0097] The element extraction module 402 is used to extract elements from each frame of the front-end page video of the user visiting the front-end page, and obtain the visual elements of each frame of the page. For detailed description, please refer to the relevant description corresponding to the above method embodiment, which will not be repeated here.

[0098] The page classification module 403 is used to classify each frame page based on the visual elements of each frame page to obtain multiple frame pages corresponding to multiple page categories. For detailed description, please refer to the relevant description corresponding to the above method embodiment, which will not be repeated here.

[0099] The anomaly detection module 404 is used to perform anomaly detection on multiple frame pages corresponding to each page category based on the page element features of each page category. For detailed description, please refer to the relevant description corresponding to the above method embodiment, which will not be repeated here.

[0100] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820 and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the front-end page abnormality detection method, which includes: obtaining a video of a user accessing a front-end page to be detected after privacy protection processing; extracting elements from each frame page in the video of the user accessing the front-end page to obtain visual elements of each frame page; classifying each frame page based on the visual elements of each frame page to obtain multiple frame pages corresponding to multiple page categories; and performing abnormality detection on multiple frame pages corresponding to each page category based on the page element features of each page category.

[0101] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0102] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the front-end page anomaly detection method provided by the above methods, which method includes: obtaining a video of a user to be detected accessing the front-end page after privacy protection processing; extracting elements of each frame page in the video of the user accessing the front-end page to obtain visual elements of each frame page; classifying each frame page based on the visual elements of each frame page to obtain multiple frame pages corresponding to multiple page categories; and performing anomaly detection on multiple frame pages corresponding to each page category based on page element features of each page category.

[0103] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the front-end page anomaly detection method provided by the above-mentioned methods, the method comprising: obtaining a video of a user to be detected accessing the front-end page after privacy protection processing; extracting elements of each frame page in the video of the user accessing the front-end page to obtain visual elements of each frame page; classifying each frame page based on the visual elements of each frame page to obtain multiple frame pages corresponding to multiple page categories; and performing anomaly detection on multiple frame pages corresponding to each page category based on page element features of each page category.

[0104] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0105] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A front-end page anomaly detection method, characterized in that: include: Obtain the video of the user to be detected accessing the front-end page after privacy protection processing; Extracting elements from each frame of the front-end page in the video of the user visiting the front-end page to obtain visual elements of each frame of the page; The extracting elements of each frame of the page in the video of the user visiting the front-end page to obtain the visual elements of each frame of the page includes: Decomposing each frame of the page in the video of the page visited by the user into elements; After the elements are disassembled, each frame page is constructed into a first frame page containing a specific element and a second frame page not containing the specific element; Rendering the first frame page and the second frame page into the browser, and obtaining corresponding rendered images respectively; Determine the difference between each frame page based on the rendered images corresponding to the first frame page and the second frame page respectively; Determining the visual element based on the difference of each frame page; Classifying each frame page based on the visual elements of each frame page to obtain a plurality of frame pages corresponding to a plurality of page categories; The classifying each frame page based on the visual elements of each frame page to obtain a plurality of frame pages corresponding to a plurality of page categories respectively includes: Based on preset page key element features and page jump features, identifying page key elements in the visual elements; Classifying each frame page based on the page key elements to obtain a plurality of frame pages corresponding to a plurality of page categories; Based on the page element features of each page category, anomaly detection is performed on multiple frame pages corresponding to each page category.

2. The method according to claim 1, characterized in that The step of obtaining the video of the user accessing the front-end page to be detected after the privacy protection processing includes: Obtain front-end page data and user access behavior data; Performing privacy shielding processing on the front-end page data to obtain processed front-end page data; Restoring the processed front-end page data into an HTML page sequence based on the user access behavior data; The HTML page sequence is rendered to obtain a video of the user accessing the page to be detected.

3. The method according to claim 2, characterized in that The performing privacy shielding processing on the front-end page data to obtain processed front-end page data includes: The user privacy information in the front-end page data is shielded by business rules, and the front-end page containing the privacy information is coded to obtain the processed front-end page data.

4. The method according to any one of claims 1 to 3, characterized in that: The method of performing abnormality detection on a plurality of frame pages corresponding to each page category based on the page element features of each page category includes: Based on the page categories, determining a plurality of target key elements contained in each page category; Based on the multiple target key elements contained in each page category, anomaly detection is performed on the multiple frame pages corresponding to each page category, wherein the anomaly detection includes missing, occlusion, and position error of the target key element.

5. The method according to claim 4, characterized in that The method further comprises: The Markov chain model is used to calculate the jump probability of each category page to other category pages; The HTML page sequences whose jump probability is lower than a preset threshold are recorded.

6. A front-end page anomaly detection device, characterized in that: include: An acquisition module is used to acquire the video of the user accessing the front-end page to be detected after the privacy protection processing; An element extraction module is used to extract elements from each frame of the front-end page video of the user visiting the front-end page to obtain visual elements of each frame of the page; The extracting elements of each frame of the page in the video of the user visiting the front-end page to obtain the visual elements of each frame of the page includes: decomposing the elements of each frame of the page in the video of the user visiting the page; constructing each frame of the page after the element decomposition into a first frame of the page containing a specific element and a second frame of the page not containing the specific element; rendering the first frame of the page and the second frame of the page to the browser, and obtaining corresponding rendering images respectively; determining the difference between each frame of the page based on the rendering images corresponding to the first frame of the page and the second frame of the page respectively; and determining the visual elements based on the difference between each frame of the page; A page classification module is used to classify each frame page based on the visual elements of each frame page to obtain a plurality of frame pages corresponding to a plurality of page categories; the classifying each frame page based on the visual elements of each frame page to obtain a plurality of frame pages corresponding to a plurality of page categories includes: identifying a page key element in the visual element based on preset page key element features and page jump features; classifying each frame page based on the page key element to obtain a plurality of frame pages corresponding to a plurality of page categories; The anomaly detection module is used to perform anomaly detection on multiple frame pages corresponding to each page category based on the page element features of each page category.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the front-end page anomaly detection method as described in any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the front-end page anomaly detection method as described in any one of claims 1 to 5 is implemented.

Images