WO2022161047A1 - 页面的埋点分析方法和装置 - Google Patents

页面的埋点分析方法和装置 Download PDF

Info

Publication number
WO2022161047A1
WO2022161047A1 PCT/CN2021/140696 CN2021140696W WO2022161047A1 WO 2022161047 A1 WO2022161047 A1 WO 2022161047A1 CN 2021140696 W CN2021140696 W CN 2021140696W WO 2022161047 A1 WO2022161047 A1 WO 2022161047A1
Authority
WO
WIPO (PCT)
Prior art keywords
real
time
buried point
screenshots
page
Prior art date
Application number
PCT/CN2021/140696
Other languages
English (en)
French (fr)
Inventor
张红
Original Assignee
北京沃东天骏信息技术有限公司
北京京东世纪贸易有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京沃东天骏信息技术有限公司, 北京京东世纪贸易有限公司 filed Critical 北京沃东天骏信息技术有限公司
Publication of WO2022161047A1 publication Critical patent/WO2022161047A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures

Definitions

  • the present disclosure relates to the field of computer technology, and in particular, to a method for analyzing a buried point of a page, a device for analyzing a buried point of a page, and a non-volatile computer-readable storage medium.
  • Buried point analysis is a data collection method for website analysis.
  • the collected data is used to track page usage in order to further optimize products or provide operational data support.
  • the analysis of buried points includes the number of visits, the number of visitors, the length of stay, the number of page views, and the bounce rate.
  • R&D personnel add embedded code in the process of developing page code in advance to realize embedded analysis.
  • a method for analyzing embedded points of a page including: taking screenshots of the page where the user is currently located, and generating multiple real-time screenshots; calculating the difference between the multiple real-time screenshots and the embedded image of each sample For the first similarity, the embedded image of each sample is generated by taking a screenshot of the changed area in the page in advance; the real-time screenshot with the first similarity greater than the first threshold is determined as the target embedded image; according to the target embedded image, Carry out page analysis.
  • calculating the first similarity between the multiple real-time screenshots and the embedded point images of each sample includes: calculating a second similarity between the multiple real-time screenshots; The screenshot is determined as a candidate buried point image; the first similarity between each candidate buried point image and each sample buried point image is calculated.
  • calculating the second similarity between the multiple real-time screenshots includes: calculating image features of each real-time screenshot; and calculating the second similarity according to the image features of each real-time screenshot.
  • calculating the image features of each real-time screenshot includes: calculating the pixel mean value of each row in each real-time screenshot; and calculating the pixel variance of each real-time screenshot as the image feature according to the pixel mean value.
  • taking a screenshot of the page where the user is currently located, and generating multiple real-time screenshots includes: obtaining the DOM (Document Object Model) of the page; converting the DOM of the page into SVG (Scalable Vector Graphics, scalable). vector graphics) image, generate a real-time screenshot in the Canvas (canvas).
  • DOM Document Object Model
  • SVG Scalable Vector Graphics, scalable. vector graphics
  • taking a screenshot of the page where the user is currently located, and generating multiple real-time screenshots includes: in response to the user entering the page, taking a regular screenshot of the page to generate multiple real-time screenshots.
  • the multiple real-time screenshots and the embedded image of each sample are stored in a key-value pair format by using form data; calculating the first similarity between the multiple real-time screenshots and the embedded image of each sample includes: according to the key Value pair format, call multiple real-time screenshots and buried point images of each sample to calculate the first similarity.
  • calculating the first similarity between the plurality of real-time screenshots and the embedded image of each sample includes: comparing the size of the real-time screenshot and the embedded image of the sample; when the real-time screenshot is larger than the size of the embedded image of the sample Under the circumstance, the real-time screenshot is reduced, so that the size of the real-time screenshot is the same as the size of the sample buried point image; in the case that the sample buried point image is larger than the real-time screenshot, the sample buried point image is reduced. processing, so that the size of the sample embedded point image is the same as the size of the real-time screenshot; calculating the first similarity between the sample embedded point image and the real-time screenshot.
  • the changed area in the page includes at least one of the area changed by the pop-up event and the area changed by the operation event.
  • calculating the first similarity between the multiple real-time screenshots and the embedded image of each sample includes: calculating image features of each real-time screenshot and image features of each embedded image of the sample; calculating the first similarity according to the image features similarity.
  • calculating the image features of each real-time screenshot and the image features of each sample embedded point image includes: calculating the pixel mean value of each row in each real-time screenshot; calculating the pixel variance of each real-time screenshot as its image feature according to the pixel mean value ; Calculate the pixel mean value of each row in each sample embedded point image; according to the pixel mean value, calculate the pixel variance of each sample embedded point image as its image feature.
  • an apparatus for analyzing embedded points of a page including: a screenshot unit, configured to take screenshots of the page where the user is currently located, and generate multiple real-time screenshots; a calculation unit, configured to calculate multiple real-time screenshots The first similarity between the screenshots and the embedded point images of each sample, and the embedded point images of each sample are generated after taking a screenshot of the changed area in the page in advance; the determining unit is used for determining the real-time similarity between the first similarity degree and the first threshold value.
  • the screenshot is determined as the target embedded point image; the analysis unit is used to perform the embedded point analysis on the page according to the target embedded point image.
  • the computing unit calculates the second similarity between multiple real-time screenshots; the determining unit determines the real-time screenshots with the second similarity less than the second threshold as candidate buried point images; the computing unit calculates each candidate buried point The first similarity between the image and the embedded image of each sample.
  • the computing unit calculates image features of each real-time screenshot; and calculates the second similarity according to the image features of each real-time screenshot.
  • the computing unit calculates the pixel mean value of each row in each real-time screenshot; and calculates the pixel variance of each real-time screenshot as an image feature according to the pixel mean value.
  • the screenshot unit acquires the DOM of the page, converts the DOM of the page into an SVG image, and generates a real-time screenshot in the Canvas.
  • the screenshot unit in response to the user entering the page, takes timed screenshots of the page to generate multiple real-time screenshots.
  • the multiple real-time screenshots and the embedded image of each sample are stored in a key-value pair format by using form data; the computing unit calls the multiple real-time screenshots and the embedded image of each sample to calculate the first similarity.
  • the computing unit compares the size of the real-time screenshot and the sample buried point image; when the real-time screenshot is larger than the sample buried point image, the real-time screenshot is reduced, so that the size of the real-time screenshot is the same as that of the sample buried point image.
  • the size of the sample buried point image is the same; when the sample buried point image is larger than the size of the real-time screenshot, the sample buried point image is reduced, so that the size of the sample buried point image is the same as that of the real-time screenshot; Calculate the first similarity between the sample buried point image and the real-time screenshot.
  • the changed area in the page includes at least one of the area changed by the pop-up event and the area changed by the operation event.
  • the calculation unit calculates the image features of each real-time screenshot and the image features of each sample embedded point image; and calculates the first similarity according to the image features.
  • the computing unit calculates the pixel mean value of each row in each real-time screenshot; according to the pixel mean value, calculates the pixel variance of each real-time screenshot as its image feature; calculates the pixel mean value of each row in each sample embedded point image; The mean value is calculated, and the pixel variance of the embedded point image of each sample is calculated as its image feature.
  • an apparatus for analyzing embedded points of a page comprising: a memory; and a processor coupled to the memory, the processor being configured to execute any one of the foregoing based on instructions stored in the memory apparatus The embedded point analysis method of the page in the embodiment.
  • a non-volatile computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, implements the method for buried point analysis of a page in any of the foregoing embodiments .
  • FIG. 1 shows a flowchart of some embodiments of the method for buried point analysis of a page of the present disclosure
  • FIG. 2 shows a flowchart of some embodiments of step 120 in FIG. 1;
  • FIG. 3 shows a flowchart of other embodiments of the buried point analysis method of the page of the present disclosure
  • FIG. 4 shows a block diagram of some embodiments of a buried point analysis apparatus for a page of the present disclosure
  • FIG. 5 shows a block diagram of other embodiments of the buried point analysis apparatus of the page of the present disclosure
  • FIG. 6 shows a block diagram of still other embodiments of buried point analysis apparatuses for pages of the present disclosure.
  • the inventors of the present disclosure found that the above-mentioned related technologies have the following problems: it is necessary to intrude into the business code to realize the development of the embedded code, and the maintenance of the later developers is difficult, resulting in high development cost of the embedded analysis.
  • the present disclosure proposes a technical solution for page buried point analysis, which can reduce development costs.
  • the following embodiment can be used to realize the buried point analyze.
  • FIG. 1 shows a flowchart of some embodiments of the buried point analysis method of the page of the present disclosure.
  • a screenshot is taken of the page where the user is currently located, and multiple real-time screenshots are generated. For example, in response to the user entering the page, take regular screenshots of the page, and generate multiple real-time screenshots.
  • real-time screenshots can be made using HTML2canvas technology. For example, it is possible to convert the fetched DOM of the page to an SVG image and then generate a live screenshot in Canvas.
  • the SVG image uses the XML (Extensible Markup Language) representation, its structure is consistent with the DOM, thereby improving the accuracy of the embedded point analysis.
  • XML Extensible Markup Language
  • the rasterizeHTML.js technology can also be used to perform real-time screenshots.
  • step 120 the first similarity between the multiple real-time screenshots and the embedded point images of each sample is calculated.
  • the embedded image of each sample is generated by taking a screenshot of the changed area in the page in advance.
  • the changed area in the page includes at least one of the area changed by the pop-up event and the area changed by the operation event.
  • a screenshot of the area in the page that needs to be analyzed for buried points may be taken in advance, and stored as a sample buried point image.
  • the sample buried point image is used to compare with multiple real-time screenshots, so as to filter out the target buried point image corresponding to the user.
  • the buried point area can be exposed for the pop-up window when the buried point analysis area is required. Before and after the pop-up window appears, you can take a screenshot of the area where the pop-up window will appear on the page as a sample buried image.
  • the area that needs to be buried point analysis can be the operation event buried point area.
  • the operation event may be a click event, a sliding event, or the like. Before and after the click operation, you can take screenshots of the area where the change effect appears on the page.
  • the real-time screenshots can be preprocessed to improve the effect and efficiency of buried point analysis.
  • real-time screenshots can be compressed.
  • the front-end JS Java Script
  • HTML5 HyperText Markup Language
  • multiple real-time screenshots and embedded point images of each sample are stored in a key-value pair format in the form of Formdata. According to the key-value pair format, call multiple real-time screenshots and embedded point images of each sample to calculate the first similarity.
  • real-time screenshots can be assembled into a set of key-value pairs sent using XML Http Request (XML HyperText Transfer Protocol Request, Extensible Markup Language Hypertext Transfer Protocol Request) through the FormData object.
  • XML Http Request XML HyperText Transfer Protocol Request, Extensible Markup Language Hypertext Transfer Protocol Request
  • the encoding type of the form can be set to multipart/form-data (multipart form data).
  • the data format of the form transmitted through FormData is the same as the data format of the form transmitted through the submit() method.
  • FormData can be used independently of the form, so the form data can be sent more flexibly.
  • the real-time screenshot and the sample buried point image may be preprocessed before calculating the first similarity.
  • the preprocessing may include image reduction processing, grayscale processing, and the like.
  • grayscale processing of real-time screenshots and sample buried point images is performed and converted into grayscale images, thereby reducing the computational complexity in the later stage.
  • the real-time screenshot is larger than the size of the sample embedded point image
  • the real-time screenshot is reduced, so that the size of the real-time screenshot is the same as the size of the sample embedded point image
  • the sample embedded point image is larger than the real-time screenshot size
  • the sample buried point image is reduced, so that the size of the sample buried point image is the same as the size of the real-time screenshot
  • the first similarity between the sample buried point image and the real-time screenshot is calculated.
  • the size of the real-time screenshots and the embedded point images of the samples can be guaranteed to be consistent, and the accuracy of the similarity calculation can be improved, thereby improving the accuracy of the embedded point analysis.
  • the image features of each real-time screenshot and the image features of each sample embedded point image are calculated; and the first similarity is calculated according to the image features.
  • image template matching or SSIM Structure SIMilarity, structural similarity
  • SSIM Structure SIMilarity, structural similarity
  • the real-time screenshots may be screened, so as to eliminate the influence of similar real-time screenshots on the analysis of buried points.
  • screening can be achieved by the embodiment in FIG. 2 .
  • FIG. 2 shows a flowchart of some embodiments of step 120 in FIG. 1 .
  • step 1210 a second similarity between multiple real-time screenshots is calculated.
  • the image features of each real-time screenshot are calculated; and the second similarity is calculated according to the image features of each real-time screenshot. For example, the pixel mean value of each row in each real-time screenshot is calculated; according to the pixel mean value, the pixel variance of each real-time screenshot is calculated as an image feature.
  • image template matching or SSIM may also be used to calculate the first similarity.
  • step 1220 the real-time screenshots with the second similarity less than the second threshold are determined as candidate buried point images.
  • step 1230 a first similarity between each candidate buried point image and each sample buried point image is calculated.
  • step 130 a real-time screenshot with the first similarity greater than the first threshold is determined as the target buried point image.
  • the target buried point image corresponding to the user can be determined for the buried point analysis according to the pre-stored sample buried point image.
  • step 140 according to the target buried point image, the page is analyzed for buried points.
  • FIG. 3 shows a flowchart of other embodiments of the buried point analysis method of the page of the present disclosure.
  • step 301 the user enters the current page and prepares to perform corresponding operations.
  • a screenshot of the area on the page that needs to be embedded point analysis can be taken in advance, and stored as a sample embedded point image, which is used to determine the user's target embedded point image through comparison later.
  • the embedded point analysis can be implemented through several steps, such as reporting user-related information, acquiring the user behavior path, and acquiring the embedded point image.
  • the step of reporting user-related information may include steps 302 and 303 .
  • step 302 the DOM of the page is loaded, and in response to the user entering the page, real-time screenshots are started.
  • screenshots may be taken using HTML2canvas technology. Convert the DOM of the page into an SVG image, and then generate the SVG image into the Canvas. In this way, since the SVG image uses the XML representation, the structure of the screenshot can be guaranteed to be consistent with the DOM of the page.
  • step 303 the real-time screenshots and the corresponding user information are stored and reported, so that the user can be analyzed for buried points. For example, you can use FormData to store and report images.
  • the real-time screenshots are assembled into a set of key-value pairs that are requested by XML Http Request through the FormData object.
  • the encoding type of the form can be set to multipart/form-data, and the data format transmitted through FormData is the same as the data format transmitted through the submit() method.
  • FormData can be used independently of the form, the form data can be sent more flexibly and conveniently.
  • the file API and Canvas technology in HTML5 can be used to realize front-end JS compression of real-time screenshots and sample embedded point images.
  • the step of obtaining the user behavior path may include steps 304 and 305 .
  • the real-time screenshot and the sample buried point image can be preprocessed.
  • the preprocessing may include image compression processing and grayscale processing.
  • image compression processing can save transmission bandwidth and storage space; grayscale processing converts images into grayscale images to reduce post-computing complexity.
  • step 305 the second similarity between the real-time screenshots is calculated, and the second score of each real-time screenshot is determined.
  • the real-time screenshot with the second higher score is determined as the candidate buried point image.
  • the second similarity is low, indicating that the user's operation behavior on the page has caused changes in the area in the page.
  • real-time screenshots with high similarity can be filtered out, and the efficiency and effect of embedded point analysis can be improved.
  • the step of acquiring buried point images may include steps 306 to 308 .
  • the real-time screenshot and the sample buried point image may be preprocessed.
  • the preprocessing can include image reduction processing, so that the size of the real-time screenshot and the embedded image of the sample remains the same.
  • step 307 the first similarity between the filtered real-time screenshot (candidate buried point image) and the sample buried point image prepared in advance is calculated, and the first score of each candidate buried point image is determined.
  • the first similarity may be calculated from the variance of the images. The smaller the difference between image variances, the higher the first similarity and the higher the first score. A high first score indicates that the candidate buried point image is similar to the sample buried point image.
  • step 308 the first image with the highest score is determined as the target buried point image.
  • step 309 according to the target buried point image, the buried point is reported.
  • the embedded point analysis is performed.
  • the image used for the embedded point analysis is determined by comparing the similarity between the real-time screenshot and the sample embedded point image. In this way, the embedded point analysis can be realized without intrusion into the business code and the participation of professional R&D personnel, thereby reducing the development cost of the embedded point analysis.
  • FIG. 4 shows a block diagram of some embodiments of a buried point analysis apparatus for a page of the present disclosure.
  • the apparatus 4 for analyzing embedded points of a page includes a screenshot unit 41 , a calculation unit 42 , a determination unit 43 and an analysis unit 44 .
  • the screenshot unit 41 takes screenshots of the page where the user is currently located, and generates multiple real-time screenshots.
  • the changed area in the page includes at least one of the area changed by the pop-up event and the area changed by the operation event.
  • the screenshot unit 41 obtains the DOM of the page, converts the DOM of the page into an SVG image, and generates a real-time screenshot in Canvas.
  • the screenshot unit 41 in response to the user entering the page, takes a regular screenshot of the page to generate multiple real-time screenshots.
  • the calculating unit 42 calculates the first similarity between the multiple real-time screenshots and the embedded point images of each sample.
  • the embedded image of each sample is generated by taking a screenshot of the changed area in the page in advance.
  • the calculating unit 42 calculates image features of each real-time screenshot; and calculates the second similarity according to the image features of each real-time screenshot.
  • the calculation unit 42 calculates the pixel mean value of each row in each real-time screenshot; according to the pixel mean value, calculates the pixel variance of each real-time screenshot as an image feature.
  • the multiple real-time screenshots and the embedded point images of each sample are stored in a key-value pair format in the form of form data; the computing unit 42 calls the multiple real-time screenshots and the embedded point images of each sample to calculate the first image according to the key-value pair format. a similarity.
  • the computing unit 42 compares the size of the real-time screenshot and the sample buried point image; when the real-time screenshot is larger than the sample buried point image, the real-time screenshot is reduced, so that the size of the real-time screenshot is The size of the embedded point image of the sample is the same; when the size of the embedded point image of the sample is larger than that of the real-time screenshot, the size of the embedded point image of the sample is reduced, so that the size of the embedded point image of the sample is the same as the size of the real-time screenshot ; Calculate the first similarity between the sample buried point image and the real-time screenshot.
  • the calculating unit 42 calculates the image features of each real-time screenshot and the image features of each sample embedded point image; and calculates the first similarity according to the image features.
  • the calculation unit 42 calculates the pixel mean value of each row in each real-time screenshot; according to the pixel mean value, calculates the pixel variance of each real-time screenshot as its image feature; calculates the pixel mean value of each row in each sample buried point image; The pixel mean value is calculated, and the pixel variance of the embedded point image of each sample is calculated as its image feature.
  • the determining unit 43 determines the real-time screenshot with the first similarity greater than the first threshold as the target buried point image.
  • the calculation unit 42 calculates the second similarity between multiple real-time screenshots; the determination unit 43 determines the real-time screenshots whose second similarity is less than the second threshold as candidate buried point images; the calculation unit 42 calculates each The first similarity between the candidate buried point images and each sample buried point image.
  • the analyzing unit 44 performs an embedded point analysis on the page according to the target embedded point image.
  • FIG. 5 shows a block diagram of other embodiments of buried point analysis apparatuses for pages of the present disclosure.
  • the embedded point analysis device 5 of the page of this embodiment includes: a memory 51 and a processor 52 coupled to the memory 51 , and the processor 52 is configured to execute the present invention based on the instructions stored in the memory 51 .
  • the memory 51 may include, for example, a system memory, a fixed non-volatile storage medium, and the like.
  • the system memory stores, for example, an operating system, an application program, a boot loader Boot Loader, a database, and other programs.
  • FIG. 6 shows a block diagram of still other embodiments of buried point analysis apparatuses for pages of the present disclosure.
  • the embedded point analysis device 6 of the page of this embodiment includes: a memory 610 and a processor 620 coupled to the memory 610 , and the processor 620 is configured to execute the foregoing based on the instructions stored in the memory 610 .
  • Memory 610 may include, for example, system memory, fixed non-volatile storage media, and the like.
  • the system memory stores, for example, an operating system, an application program, a boot loader, and other programs.
  • the embedded point analysis device 6 of the page may further include an input and output interface 630, a network interface 640, a storage interface 650, and the like. These interfaces 630 , 640 , 650 and the memory 610 and the processor 620 may be connected, for example, through a bus 660 .
  • the input and output interface 630 provides a connection interface for input and output devices such as a display, a mouse, a keyboard, a touch screen, a microphone, and a speaker.
  • Network interface 640 provides a connection interface for various networked devices.
  • the storage interface 650 provides a connection interface for external storage devices such as SD cards and U disks.
  • embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable non-transitory storage media having computer-usable program code embodied therein, including but not limited to disk storage, CD-ROM, optical storage, and the like.
  • the methods and systems of the present disclosure may be implemented in many ways.
  • the methods and systems of the present disclosure may be implemented in software, hardware, firmware, or any combination of software, hardware, and firmware.
  • the above order of steps for the method is for illustration only, and the steps of the method of the present disclosure are not limited to the order specifically described above unless specifically stated otherwise.
  • the present disclosure can also be implemented as programs recorded in a recording medium, the programs including machine-readable instructions for implementing methods according to the present disclosure.
  • the present disclosure also covers a recording medium storing a program for executing the method according to the present disclosure.

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Processing Or Creating Images (AREA)

Abstract

本公开涉及一种页面的埋点分析方法和装置，涉及计算机技术领域。该埋点分析方法包括：对用户当前所在的页面进行截图，生成多张实时截图；分别计算多张实时截图与样本埋点图像之间的第一相似度，样本埋点图像为预先对页面中发生变化的区域进行截图后生成；将第一相似度大于第一阈值的一张或多张实时截图，确定为一张或多张目标埋点图像；根据目标埋点图像，进行页面进行埋点分析。

Description

页面的埋点分析方法和装置
相关申请的交叉引用
本申请是以CN申请号为202110126414.9，申请日为2021年1月29日的申请为基础，并主张其优先权，该CN申请的公开内容在此作为整体引入本申请中。
技术领域
本公开涉及和计算机技术领域，特别涉及一种页面的埋点分析方法、页面的埋点分析装置和非易失性计算机可读存储介质。
背景技术
埋点分析是网站分析的一种数据采集方法。采集的数据用于跟踪页面的使用情况，以便进一步优化产品或是提供运营的数据支撑。例如，埋点分析包括访问数、访客数、停留时长、页面浏览数、跳出率等。
在相关技术中，研发人员根据产品需求，预先在开发页面代码过程中加入埋点代码，以实现埋点分析。
发明内容
根据本公开的一些实施例，提供了一种页面的埋点分析方法，包括：对用户当前所在的页面进行截图，生成多张实时截图；计算多张实时截图与各样本埋点图像之间的第一相似度，各样本埋点图像为预先对页面中发生变化的区域进行截图后生成；将第一相似度大于第一阈值的实时截图，确定为目标埋点图像；根据目标埋点图像，进行页面进行埋点分析。
在一些实施例中，计算多张实时截图与各样本埋点图像之间的第一相似度包括：计算多张实时截图之间的第二相似度；将第二相似度小于第二阈值的实时截图，确定为候选埋点图像；计算各候选埋点图像与各样本埋点图像之间的第一相似度。
在一些实施例中，计算多张实时截图之间的第二相似度包括：计算各实时截图的图像特征；根据各实时截图的图像特征，计算第二相似度。
在一些实施例中，计算各实时截图的图像特征包括：计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为图像特征。
在一些实施例中，对用户当前所在的页面进行截图，生成多张实时截图包括：获取页面的DOM(Document Object Model，文档对象模型)；将页面的DOM转换为SVG(Scalable Vector Graphics，可缩放矢量图形)图像后，在Canvas(画布)中生成实时截图。
在一些实施例中，对用户当前所在的页面进行截图，生成多张实时截图包括：响应于用户进入页面，对页面进行定时截图，生成多张实时截图。
在一些实施例中，多张实时截图和各样本埋点图像采用表单数据的方式存储为键值对格式；计算多张实时截图与各样本埋点图像之间的第一相似度包括：根据键值对格式，调用多张实时截图和各样本埋点图像计算第一相似度。
在一些实施例中，计算多张实时截图与各样本埋点图像之间的第一相似度包括：比较实时截图与样本埋点图像的尺寸大小；在实时截图比样本埋点图像的尺寸大的情况下，对该实时截图进行缩小处理，使得该实时截图的尺寸与该样本埋点图像的尺寸相同；在样本埋点图像比实时截图的尺寸大的情况下，对该样本埋点图像进行缩小处理，使得该样本埋点图像的尺寸与该实时截图的尺寸相同；计算该样本埋点图像与该实时截图之间的第一相似度。
在一些实施例中，页面中发生变化的区域包括弹窗事件导致变化的区域、操作事件导致变化的区域中的至少一项。
在一些实施例中，计算多张实时截图与各样本埋点图像之间的第一相似度包括：计算各实时截图的图像特征和各样本埋点图像的图像特征；根据图像特征，计算第一相似度。
在一些实施例中，计算各实时截图的图像特征和各样本埋点图像的图像特征包括：计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为其图像特征；计算各样本埋点图像中每一行的像素均值；根据像素均值，计算各样本埋点图像的像素方差作为其图像特征。
根据本公开的另一些实施例，提供一种页面的埋点分析装置，包括：截图单元，用于对用户当前所在的页面进行截图，生成多张实时截图；计算单元，用于计算多张实时截图与各样本埋点图像之间的第一相似度，各样本埋点图像为预先对页面中发生变化的区域进行截图后生成；确定单元，用于将第一相似度大于第一阈值的实时截图，确定为目标埋点图像；分析单元，用于根据目标埋点图像，进行页面进行埋点分析。
在一些实施例中，计算单元计算多张实时截图之间的第二相似度；确定单元将第 二相似度小于第二阈值的实时截图，确定为候选埋点图像；计算单元计算各候选埋点图像与各样本埋点图像之间的第一相似度。
在一些实施例中，计算单元计算各实时截图的图像特征；根据各实时截图的图像特征，计算第二相似度。
在一些实施例中，计算单元计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为图像特征。
在一些实施例中，截图单元获取页面的DOM，将述页面的DOM转换为SVG图像后，在Canvas中生成实时截图。
在一些实施例中，截图单元响应于用户进入页面，对页面进行定时截图，生成多张实时截图。
在一些实施例中，多张实时截图和各样本埋点图像采用表单数据的方式存储为键值对格式；计算单元根据键值对格式，调用多张实时截图和各样本埋点图像计算第一相似度。
在一些实施例中，计算单元比较实时截图与样本埋点图像的尺寸大小；在实时截图比样本埋点图像的尺寸大的情况下，对该实时截图进行缩小处理，使得该实时截图的尺寸与该样本埋点图像的尺寸相同；在样本埋点图像比实时截图的尺寸大的情况下，对该样本埋点图像进行缩小处理，使得该样本埋点图像的尺寸与该实时截图的尺寸相同；计算该样本埋点图像与该实时截图之间的第一相似度。
在一些实施例中，页面中发生变化的区域包括弹窗事件导致变化的区域、操作事件导致变化的区域中的至少一项。
在一些实施例中，计算单元计算各实时截图的图像特征和各样本埋点图像的图像特征；根据图像特征，计算第一相似度。
在一些实施例中，计算单元计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为其图像特征；计算各样本埋点图像中每一行的像素均值；根据像素均值，计算各样本埋点图像的像素方差作为其图像特征。
根据本公开的又一些实施例，提供一种页面的埋点分析装置，包括：存储器；和耦接至存储器的处理器，处理器被配置为基于存储在存储器装置中的指令，执行上述任一个实施例中的页面的埋点分析方法。
根据本公开的再一些实施例，提供一种非易失性计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现上述任一个实施例中的页面的埋点分析方 法。
附图说明
构成说明书的一部分的附图描述了本公开的实施例，并且连同说明书一起用于解释本公开的原理。
参照附图，根据下面的详细描述，可以更加清楚地理解本公开：
图1示出本公开的页面的埋点分析方法的一些实施例的流程图；
图2示出图1中步骤120的一些实施例的流程图；
图3示出本公开的页面的埋点分析方法的另一些实施例的流程图；
图4示出本公开的页面的埋点分析装置的一些实施例的框图；
图5示出本公开的页面的埋点分析装置的另一些实施例的框图；
图6示出本公开的页面的埋点分析装置的又一些实施例的框图。
具体实施方式
现在将参照附图来详细描述本公开的各种示例性实施例。应注意到：除非另外具体说明，否则在这些实施例中阐述的部件和步骤的相对布置、数字表达式和数值不限制本公开的范围。
同时，应当明白，为了便于描述，附图中所示出的各个部分的尺寸并不是按照实际的比例关系绘制的。
以下对至少一个示例性实施例的描述实际上仅仅是说明性的，决不作为对本公开及其应用或使用的任何限制。
对于相关领域普通技术人员已知的技术、方法和设备可能不作详细讨论，但在适当情况下，技术、方法和设备应当被视为授权说明书的一部分。
在这里示出和讨论的所有示例中，任何具体值应被解释为仅仅是示例性的，而不是作为限制。因此，示例性实施例的其它示例可以具有不同的值。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步讨论。
本公开的发明人发现上述相关技术中存在如下问题：需要入侵到业务代码实现埋点代码的开发，且后期开发人员维护困难，导致埋点分析的开发成本高。
鉴于此，本公开提出了一种页面的埋点分析技术方案，能够降低开发成本。
如前所述，针对研发人员需要侵入到业务代码才能上报埋点，且只能由研发人员来完成埋点处理，导致的开发成本较高的技术问题，可以采用下面的实施例来实现埋点分析。
图1示出本公开的页面的埋点分析方法的一些实施例的流程图。
如图1所示，在步骤110中，对用户当前所在的页面进行截图，生成多张实时截图。例如，响应于用户进入页面，对页面进行定时截图，生成多张实时截图。
在一些实施例中，可以利用HTML2canvas技术进行实时截图。例如，可以将获取的页面的DOM转换为SVG图像，然后在Canvas中生成实时截图。
这样，由于SVG图像使用的是XML(Extensible Markup Language，可扩展标记语言)表示方式，其结构与DOM保持一致，从而提高了埋点分析的准确度。
在一些实施例中，也可以利用rasterizeHTML.js技术进行实时截图。
在步骤120中，计算多张实时截图与各样本埋点图像之间的第一相似度。
在一些实施例中，各样本埋点图像为预先对该页面中发生变化的区域进行截图后生成。例如，页面中发生变化的区域包括弹窗事件导致变化的区域、操作事件导致变化的区域中的至少一项。
在一些实施例中，可以预先对页面中需要进行埋点分析的区域进行截图，并存储为样本埋点图像。样本埋点图像用于与多张实时截图进行对比，从而筛选出与用户对应的目标埋点图像。
例如，需要进行埋点分析区域可以为弹窗曝光埋点区域。可以在弹窗出现前后，对页面中会产生弹窗的区域进行截图作为样本埋点图像。
例如，需要进行埋点分析区域可以为操作事件埋点区域。操作事件可以为点击事件、滑动事件等。可以在点击操作前后，对页面中出现变化效果的区域进行截图。
在一些实施例中，可以对实时截图进行预处理，以提高埋点分析的效果和效率。例如，可以对实时截图进行压缩处理。
例如，可以通过HTML5(HyperText Markup Language，超文本标记语言)中的file API(file Application Programming Interface，文件应用程序接口)和Canvas技术实现实时截图的前端JS(Java Script)压缩。
这样，可以节省提交埋点的用户的带宽，并节省存储空间。
在一些实施例中，多张实时截图和各样本埋点图像采用Formdata(表单数据)的方式存储为键值对格式。根据键值对格式，调用多张实时截图和各样本埋点图像计算 第一相似度。
这样，采用FormData的键值对格式提交和存储图像能够实现表单数据的序列化，从而减少表单元素的拼接，提高工作效率。
例如，可以通过FormData对象，将实时截图组装为一组利用XML Http Request(XML HyperText Transfer Protocol Request，可扩展标记语言超文本传输协议请求)发送的键值对。
可以将表单的编码类型设置为multipart/form-data(多部分表单数据)。通过FormData传输表单的数据格式与通过submit()方法传输表单的数据格式一致。
这样，FormData可以独立于表单使用，因此可以更灵活的发送表单数据。
在一些实施例中，在计算第一相似度之前可以对实时截图、样本埋点图像进行预处理。例如，预处理可以包括图像缩小处理、灰度化处理等。
例如，对实时截图、样本埋点图像进行灰度化处理，转换为灰度图，从而减少后期计算复杂度。
例如，比较实时截图与样本埋点图像的尺寸大小。在实时截图比样本埋点图像的尺寸大的情况下，对该实时截图进行缩小处理，使得该实时截图的尺寸与该样本埋点图像的尺寸相同；在样本埋点图像比实时截图的尺寸大的情况下，对该样本埋点图像进行缩小处理，使得该样本埋点图像的尺寸与该实时截图的尺寸相同；计算该样本埋点图像与该实时截图之间的第一相似度。
这样，可以保证实时截图、样本埋点图像的大小一致，提高相似度计算的准确性，从而提高埋点分析的准确性。
在一些实施例中，计算各实时截图的图像特征和各样本埋点图像的图像特征；根据图像特征，计算第一相似度。
例如，计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为其图像特征；计算各样本埋点图像中每一行的像素均值；根据像素均值，计算各样本埋点图像的像素方差作为其图像特征。
例如，也可以利用图像模版匹配或者SSIM(Structural SIMilarity，结构相似性)计算第一相似度。
在一些实施例中，在计算第一相似度之前，可以对实时截图进行筛选，从而消除相似的实时截图对埋点分析的影响。例如，可以通过图2中的实施例实现筛选。
图2示出图1中步骤120的一些实施例的流程图。
如图2所示，在步骤1210中，计算多张实时截图之间的第二相似度。
在一些实施例中，计算各实时截图的图像特征；根据各实时截图的图像特征，计算第二相似度。例如，计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为图像特征。
在一些实施例中，也可以利用图像模版匹配或者SSIM计算第一相似度。
在步骤1220中，将第二相似度小于第二阈值的实时截图，确定为候选埋点图像。
在步骤1230中，计算各候选埋点图像与各样本埋点图像之间的第一相似度。
这样，可以筛选掉相似程度较高的实时截图，只保留区别较大的实时截图作为候选埋点图像，从而提高了埋点分析的效率和效果。在计算了第一相似度之后，可以利用图1中的剩余步骤进行埋点分析。
在步骤130中，将第一相似度大于第一阈值的实时截图，确定为目标埋点图像。这样，无需专业人员参与和重新开发代码，即可根据预先存储的样本埋点图像，确定用户对应的目标埋点图像用于埋点分析。
在步骤140中，根据目标埋点图像，进行页面进行埋点分析。
图3示出本公开的页面的埋点分析方法的另一些实施例的流程图。
如图3所示，在步骤301中，用户进入当前的页面，准备进行相应的操作。例如，可以预先对页面中需要埋点分析的区域进行截图，存储为样本埋点图像，用于后期通过比对确定用户的目标埋点图像。
在一些实施例中，在用户进入当前的页面后，可以通过上报用户相关信息、获取用户行为路径、获取埋点图像等几个步骤实现埋点分析。
在一些实施例中，上报用户相关信息步骤可以包括步骤302、步骤303。
在步骤302中，加载页面的DOM，响应于用户进入页面，开始进行实时截图。
在一些实施例中，可以利用HTML2canvas技术进行截图。将页面的DOM转成SVG图像，然后将SVG图像生成到Canvas中。这样，由于SVG图像使用的是XML表示形式，可以保证截图的结构与页面的DOM一致。
在步骤303中，对实时截图和相应的用户信息进行存储和上报，用于对该用户进行埋点分析。例如，可以利用FormData实现图像的存储和上报。
在一些实施例中，通过FormData对象将实时截图组装为一组用XML Http Request发送请求的键值对。可以将表单的编码类型设置为multipart/form-data，而且通过FormData传输的数据格式与通过submit()方法传输的数据格式一致。
这样，由于FormData可独立于表单使用，因此可以更灵活方便地发送表单数据。
在一些实施例中，在进行存储和上报之前，可以利用HTML5中的file API和Canvas技术，实现实时截图和样本埋点图像的前端JS压缩。
在一些实施例中，获取用户行为路径步骤可以包括步骤304、步骤305。
在步骤304中，可以对实时截图和样本埋点图像进行预处理。例如，预处理可以包括图像压缩处理、灰度化处理。
例如，图像压缩处理可以节省传输带宽和存储空间；灰度化处理将图像转换为灰度图，以减少后期计算复杂度。
在步骤305中，计算实时截图之间的第二相似度，并确定各实时截图的第二得分。
在一些实施例中，分别计算各实时截图每行像素点的平均值υ，每一个υ对应一行图像的特征；根据各实时截图中各行像素点的υ，计算各实时截图的方差σ作为图像特征，用于示每行像素的波动情况；根据各实时截图的方差之间的差异，计算各实时截图的第二得分。例如，可以根据C＝1/σ计算第二得分，用于表征图像之间的相似程度。
例如，图像方差之间的差异越大，第二相似度越低，第二得分越高。将第二得分较高的实时截图确定为候选埋点图像。
第二相似度低表示用户对页面的操作行为导致页面中的区域发生了变化。这样，可以过滤掉相似度高的实时截图，提高埋点分析的效率和效果。
在一些实施例中，获取埋点图像步骤可以包括步骤306～步骤308。
在步骤306中，可以对实时截图和样本埋点图像进行预处理。例如，预处理可以包括图像缩小处理，使得实时截图和样本埋点图像的大小保持一致。
在步骤307中，计算过滤后的实时截图(候选埋点图像)与事先准备的样本埋点图像的第一相似度，并确定各候选埋点图像的第一得分。
在一些实施例中，可以通过图像的方差计算第一相似度。图像方差之间的差异越小，第一相似度越高，第一得分越高。第一得分高说明候选埋点图像与样本埋点图像比较相似。
在步骤308中，将第一得分高的图像，确定为目标埋点图像。
在步骤309中，根据目标埋点图像，进行埋点上报。通过分析上报的截图对应的用户的操作信息、用户的身份信息等用户相关信息，进行埋点分析。
在上述实施例中，通过对比实时截图与样本埋点图像之间的相似度，确定用于埋点分析的图像。这样，无需入侵到业务代码和专业的研发人员参与即可实现埋点分析， 从而降低了埋点分析的开发成本。
图4示出本公开的页面的埋点分析装置的一些实施例的框图。
如图4所示，页面的埋点分析装置4包括截图单元41、计算单元42、确定单元43和分析单元44。
截图单元41对用户当前所在的页面进行截图，生成多张实时截图。
在一些实施例中，页面中发生变化的区域包括弹窗事件导致变化的区域、操作事件导致变化的区域中的至少一项。
在一些实施例中，截图单元41获取页面的DOM，将述页面的DOM转换为SVG图像后，在Canvas中生成实时截图。
在一些实施例中，截图单元41响应于用户进入页面，对页面进行定时截图，生成多张实时截图。
计算单元42计算多张实时截图与各样本埋点图像之间的第一相似度。各样本埋点图像为预先对页面中发生变化的区域进行截图后生成。
在一些实施例中，计算单元42计算各实时截图的图像特征；根据各实时截图的图像特征，计算第二相似度。
在一些实施例中，计算单元42计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为图像特征。
在一些实施例中，多张实时截图和各样本埋点图像采用表单数据的方式存储为键值对格式；计算单元42根据键值对格式，调用多张实时截图和各样本埋点图像计算第一相似度。
在一些实施例中，计算单元42比较实时截图与样本埋点图像的尺寸大小；在实时截图比样本埋点图像的尺寸大的情况下，对该实时截图进行缩小处理，使得该实时截图的尺寸与该样本埋点图像的尺寸相同；在样本埋点图像比实时截图的尺寸大的情况下，对该样本埋点图像进行缩小处理，使得该样本埋点图像的尺寸与该实时截图的尺寸相同；计算该样本埋点图像与该实时截图之间的第一相似度。
在一些实施例中，计算单元42计算各实时截图的图像特征和各样本埋点图像的图像特征；根据图像特征，计算第一相似度。
在一些实施例中，计算单元42计算各实时截图中每一行的像素均值；根据像素均值，计算各实时截图的像素方差作为其图像特征；计算各样本埋点图像中每一行的像素均值；根据像素均值，计算各样本埋点图像的像素方差作为其图像特征。
确定单元43将第一相似度大于第一阈值的实时截图，确定为目标埋点图像。
在一些实施例中，计算单元42计算多张实时截图之间的第二相似度；确定单元43将第二相似度小于第二阈值的实时截图，确定为候选埋点图像；计算单元42计算各候选埋点图像与各样本埋点图像之间的第一相似度。
分析单元44根据目标埋点图像，进行页面进行埋点分析。
图5示出本公开的页面的埋点分析装置的另一些实施例的框图。
如图5所示，该实施例的页面的埋点分析装置5包括：存储器51以及耦接至该存储器51的处理器52，处理器52被配置为基于存储在存储器51中的指令，执行本公开中任意一个实施例中的页面的埋点分析方法。
其中，存储器51例如可以包括系统存储器、固定非易失性存储介质等。系统存储器例如存储有操作系统、应用程序、引导装载程序Boot Loader、数据库以及其他程序等。
图6示出本公开的页面的埋点分析装置的又一些实施例的框图。
如图6所示，该实施例的页面的埋点分析装置6包括：存储器610以及耦接至该存储器610的处理器620，处理器620被配置为基于存储在存储器610中的指令，执行前述任意一个实施例中的页面的埋点分析方法。
存储器610例如可以包括系统存储器、固定非易失性存储介质等。系统存储器例如存储有操作系统、应用程序、引导装载程序Boot Loader以及其他程序等。
页面的埋点分析装置6还可以包括输入输出接口630、网络接口640、存储接口650等。这些接口630、640、650以及存储器610和处理器620之间例如可以通过总线660连接。其中，输入输出接口630为显示器、鼠标、键盘、触摸屏、麦克、音箱等输入输出设备提供连接接口。网络接口640为各种联网设备提供连接接口。存储接口650为SD卡、U盘等外置存储设备提供连接接口。
本领域内的技术人员应当明白，本公开的实施例可提供为方法、系统、或计算机程序产品。因此，本公开可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本公开可采用在一个或多个其中包含有计算机可用程序代码的计算机可用非瞬时性存储介质包括但不限于磁盘存储器、CD-ROM、光学存储器等上实施的计算机程序产品的形式。
至此，已经详细描述了根据本公开的页面的埋点分析方法、页面的埋点分析装置和非易失性计算机可读存储介质。为了避免遮蔽本公开的构思，没有描述本领域所公 知的一些细节。本领域技术人员根据上面的描述，完全可以明白如何实施这里公开的技术方案。
可能以许多方式来实现本公开的方法和系统。例如，可通过软件、硬件、固件或者软件、硬件、固件的任何组合来实现本公开的方法和系统。用于方法的步骤的上述顺序仅是为了进行说明，本公开的方法的步骤不限于以上具体描述的顺序，除非以其它方式特别说明。此外，在一些实施例中，还可将本公开实施为记录在记录介质中的程序，这些程序包括用于实现根据本公开的方法的机器可读指令。因而，本公开还覆盖存储用于执行根据本公开的方法的程序的记录介质。
虽然已经通过示例对本公开的一些特定实施例进行了详细说明，但是本领域的技术人员应该理解，以上示例仅是为了进行说明，而不是为了限制本公开的范围。本领域的技术人员应该理解，可在不脱离本公开的范围和精神的情况下，对以上实施例进行修改。本公开的范围由所附权利要求来限定。

Claims (16)

  1. 一种页面的埋点分析方法，包括：
    对用户当前所在的页面进行截图，生成多张实时截图；
    分别计算所述多张实时截图与样本埋点图像之间的第一相似度，所述样本埋点图像为预先对所述页面中发生变化的区域进行截图后生成；
    将第一相似度大于第一阈值的一张或多张实时截图，确定为一张或多张目标埋点图像；
    根据所述目标埋点图像，进行所述页面进行埋点分析。
  2. 根据权利要求1所述的埋点分析方法，其中，所述分别计算所述多张实时截图与样本埋点图像之间的第一相似度包括：
    计算所述多张实时截图之间的第二相似度；
    将第二相似度小于第二阈值的一张或多张实时截图，确定为一张或多张候选埋点图像；
    计算各候选埋点图像与样本埋点图像之间的所述第一相似度。
  3. 根据权利要求2所述的埋点分析方法，其中，所述计算所述多张实时截图之间的第二相似度包括：
    计算各实时截图的图像特征；
    根据所述各实时截图的图像特征，计算所述第二相似度。
  4. 根据权利要求3所述的埋点分析方法，其中，所述计算各实时截图的图像特征包括：
    计算所述各实时截图中每一行的像素均值；
    根据所述像素均值，计算所述各实时截图的像素方差作为所述图像特征。
  5. 根据权利要求1所述的埋点分析方法，其中，所述对用户当前所在的页面进行截图，生成多张实时截图包括：
    获取所述页面的文档对象模型DOM；
    将所述页面的DOM转换为可缩放矢量图形SVG图像后，在画布Canvas中生成实时截图。
  6. 根据权利要求1所述的埋点分析方法，其中，所述对用户当前所在的页面进行截图，生成多张实时截图包括：
    响应于所述用户进入所述页面，对所述页面进行定时截图，生成所述多张实时截图。
  7. 根据权利要求1所述的埋点分析方法，其中，
    所述多张实时截图和所述样本埋点图像采用表单数据的方式存储为键值对格式；
    所述分别计算所述多张实时截图与样本埋点图像之间的第一相似度包括：
    根据所述键值对格式，调用所述多张实时截图和所述样本埋点图像计算所述第一相似度。
  8. 根据权利要求1所述的埋点分析方法，其中，所述分别计算所述多张实时截图与样本埋点图像之间的第一相似度包括：
    比较实时截图与样本埋点图像的尺寸大小；
    在所述实时截图比所述样本埋点图像的尺寸大的情况下，对所述实时截图进行缩小处理，使得所述实时截图的尺寸与所述样本埋点图像的尺寸相同；
    在所述样本埋点图像比所述实时截图的尺寸大的情况下，对所述样本埋点图像进行缩小处理，使得所述样本埋点图像的尺寸与所述实时截图的尺寸相同；
    计算所述样本埋点图像与所述实时截图之间的第一相似度。
  9. 根据权利要求1-8任一项所述的埋点分析方法，其中，
    所述页面中发生变化的区域包括弹窗事件导致变化的区域、操作事件导致变化的区域中的至少一项。
  10. 根据权利要求1-8任一项所述的埋点分析方法，其中，所述分别计算所述多张实时截图与样本埋点图像之间的第一相似度包括：
    分别计算各实时截图的图像特征和所述样本埋点图像的图像特征；
    根据所述图像特征，计算所述第一相似度。
  11. 根据权利要求10所述的埋点分析方法，其中，所述分别计算各实时截图的图像特征和所述样本埋点图像的图像特征包括：
    计算所述各实时截图中每一行的像素均值；
    根据像素均值，计算所述各实时截图的像素方差作为其图像特征；
    计算所述样本埋点图像中每一行的像素均值；
    根据所述像素均值，计算所述样本埋点图像的像素方差作为其图像特征。
  12. 一种页面的埋点分析装置，包括：
    截图单元，用于对用户当前所在的页面进行截图，生成多张实时截图；
    计算单元，用于分别计算所述多张实时截图与样本埋点图像之间的第一相似度，所述样本埋点图像为预先对所述页面中发生变化的区域进行截图后生成；
    确定单元，用于将第一相似度大于第一阈值的一张或多张实时截图，确定为一张或多张目标埋点图像；
    分析单元，用于根据所述目标埋点图像，进行所述页面进行埋点分析。
  13. 根据权利要求12所述的埋点分析装置，其中，
    所述计算单元计算所述多张实时截图之间的第二相似度；
    所述确定单元将第二相似度小于第二阈值的一张或多张实时截图，确定为一张或多张候选埋点图像；
    所述计算单元计算各候选埋点图像与样本埋点图像之间的所述第一相似度。
  14. 根据权利要求12所述的埋点分析装置，其中，
    所述截图单元获取所述页面的文档对象模型DOM，将述页面的DOM转换为可缩放矢量图形SVG图像后，在画布Canvas中生成实时截图。
  15. 一种页面的埋点分析装置，包括：
    存储器；和
    耦接至所述存储器的处理器，所述处理器被配置为基于存储在所述存储器中的指令，执行权利要求1-11任一项所述的页面的埋点分析方法。
  16. 一种非易失性计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现权利要求1-11任一项所述的页面的埋点分析方法。
PCT/CN2021/140696 2021-01-29 2021-12-23 页面的埋点分析方法和装置 WO2022161047A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110126414.9 2021-01-29
CN202110126414.9A CN113762312A (zh) 2021-01-29 2021-01-29 页面的埋点分析方法和装置

Publications (1)

Publication Number Publication Date
WO2022161047A1 true WO2022161047A1 (zh) 2022-08-04

Family

ID=78786528

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/140696 WO2022161047A1 (zh) 2021-01-29 2021-12-23 页面的埋点分析方法和装置

Country Status (2)

Country Link
CN (1) CN113762312A (zh)
WO (1) WO2022161047A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113762312A (zh) * 2021-01-29 2021-12-07 北京沃东天骏信息技术有限公司 页面的埋点分析方法和装置
CN115514678B (zh) * 2022-09-23 2023-09-26 四川新网银行股份有限公司 一种互联网金融业务的连续性监控方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017008744A1 (zh) * 2015-07-13 2017-01-19 西安中兴新软件有限责任公司 一种获取自动化脚本、自动化脚本的应用方法及装置
CN109766256A (zh) * 2018-12-21 2019-05-17 中国平安财产保险股份有限公司 应用程序中h5页面性能测试方法、装置和计算机设备
CN111314721A (zh) * 2020-02-11 2020-06-19 北京达佳互联信息技术有限公司 一种异常直播的确定方法、装置、设备及介质
CN113762312A (zh) * 2021-01-29 2021-12-07 北京沃东天骏信息技术有限公司 页面的埋点分析方法和装置

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104899146B (zh) * 2015-06-19 2018-04-24 安一恒通（北京）科技有限公司 基于图像匹配技术的软件稳定性测试方法和装置
CN107995266A (zh) * 2017-11-22 2018-05-04 平安科技（深圳）有限公司 埋点数据处理方法、装置、计算机设备和存储介质
CN110704772A (zh) * 2018-06-22 2020-01-17 北京京东尚科信息技术有限公司 页面异常监控方法、系统、装置、电子设备及计算机可读介质
CN110099283A (zh) * 2019-05-09 2019-08-06 广州虎牙信息科技有限公司 信息推送方法、装置、设备和存储介质
CN111506802A (zh) * 2020-03-16 2020-08-07 中国平安人寿保险股份有限公司 一种用户信息修正方法、装置、计算机设备及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017008744A1 (zh) * 2015-07-13 2017-01-19 西安中兴新软件有限责任公司 一种获取自动化脚本、自动化脚本的应用方法及装置
CN109766256A (zh) * 2018-12-21 2019-05-17 中国平安财产保险股份有限公司 应用程序中h5页面性能测试方法、装置和计算机设备
CN111314721A (zh) * 2020-02-11 2020-06-19 北京达佳互联信息技术有限公司 一种异常直播的确定方法、装置、设备及介质
CN113762312A (zh) * 2021-01-29 2021-12-07 北京沃东天骏信息技术有限公司 页面的埋点分析方法和装置

Also Published As

Publication number Publication date
CN113762312A (zh) 2021-12-07

Similar Documents

Publication Publication Date Title
AU2020200058B2 (en) Image quality assessment and improvement for performing optical character recognition
US8549395B2 (en) Method and system for transforming an integrated webpage
WO2022161047A1 (zh) 页面的埋点分析方法和装置
US9300672B2 (en) Managing user access to query results
CN105740707B (zh) 恶意文件的识别方法和装置
Lytvyn et al. System development for video stream data analyzing
US20170185913A1 (en) System and method for comparing training data with test data
CN113743607A (zh) 异常检测模型的训练方法、异常检测方法及装置
US20200225927A1 (en) Methods and systems for automating computer application tasks using application guides, markups and computer vision
WO2022105003A1 (zh) 医疗信息处理方法、装置及电子设备
CN114245232A (zh) 一种视频摘要生成方法、装置、存储介质及电子设备
CN112035205A (zh) 数据处理方法、装置、设备和存储介质
CN106250397B (zh) 一种用户行为特征的分析方法及装置
CN114139630A (zh) 姿态识别方法、装置、存储介质和电子设备
US20230394030A1 (en) Generating event logs from video streams
CN113656391A (zh) 数据检测方法及装置、存储介质及电子设备
CN113538413A (zh) 图像检测方法及装置、电子设备和存储介质
CN117633613A (zh) 跨模态的视频情感分析方法及装置、设备、存储介质
CN112989763A (zh) 数据获取方法、装置、计算机设备及存储介质
CN110674497B (zh) 一种恶意程序相似度计算的方法和装置
CN112307386A (zh) 信息监控方法、系统、电子设备及计算机可读存储介质
CN116881971A (zh) 一种敏感信息泄露检测方法、设备及存储介质
CN113673214A (zh) 信息清单的对齐方法、装置、存储介质和电子设备
JP7429374B2 (ja) 情報処理システム、情報処理方法及び情報処理プログラム
CN116881915B (zh) 文件检测方法、电子设备和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21922637

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21922637

Country of ref document: EP

Kind code of ref document: A1