WO2021036454A1 - 一种基于屏幕分割的人机相似轨迹检测方法及装置 - Google Patents

一种基于屏幕分割的人机相似轨迹检测方法及装置 Download PDF

Info

Publication number
WO2021036454A1
WO2021036454A1 PCT/CN2020/097854 CN2020097854W WO2021036454A1 WO 2021036454 A1 WO2021036454 A1 WO 2021036454A1 CN 2020097854 W CN2020097854 W CN 2020097854W WO 2021036454 A1 WO2021036454 A1 WO 2021036454A1
Authority
WO
WIPO (PCT)
Prior art keywords
trajectory
sequence
screen
behavior data
coordinate list
Prior art date
Application number
PCT/CN2020/097854
Other languages
English (en)
French (fr)
Inventor
裴银银
叶国华
房树志
温烨伟
何鸿雪
Original Assignee
苏宁云计算有限公司
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 苏宁云计算有限公司 filed Critical 苏宁云计算有限公司
Priority to CA3152854A priority Critical patent/CA3152854C/en
Publication of WO2021036454A1 publication Critical patent/WO2021036454A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0609—Buyer or seller confidence or verification
    • 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 invention relates to the technical field of e-commerce risk control, in particular to a method and device for detecting similar human-machine trajectories based on screen segmentation.
  • MD5 Message Digest Algorithm MD5
  • MD5 is called the fifth edition of message digest algorithm in Chinese. It is a hash function widely used in the field of computer security to provide integrity protection of messages. MD5 is used to ensure complete and consistent information transmission. It is one of the hash algorithms (also translated digest algorithm and hash algorithm) widely used by computers. It is mainly used to calculate data (such as Chinese characters) into another fixed length value.
  • the current man-machine trajectory is characterized by the overall similarity of the replayed trajectory images, and the local coordinates have offsets of up, down, left, and right. The exact same trajectory. Therefore, the accuracy of detection is not high.
  • embodiments of the present invention provide a method for detecting similar human-machine trajectories based on screen segmentation to overcome the overall similarity of the replayed trajectory images in the prior art, and the local coordinates have offsets of up, down, left, and right. , It is difficult to have two identical trajectories, leading to problems such as low detection accuracy.
  • the technical solution adopted by the present invention is:
  • a human-machine similar trajectory detection method based on screen segmentation includes the following steps:
  • the count value is compared with a preset threshold value, and the track sequence whose count value exceeds the preset threshold value is determined as a scalper track.
  • the method further includes:
  • An initialization sequence is generated according to the grid-like image, wherein the initialization trajectory sequence is m*n characters '0'.
  • the method further includes:
  • obtaining the trajectory sequence corresponding to the behavior data to be detected includes:
  • the method further includes:
  • the obtained new trajectory sequence is compared with the scalper trajectory, and if they are consistent, the new trajectory sequence is determined as the scalper trajectory.
  • a device for detecting similar human-machine trajectories based on screen segmentation includes:
  • a coordinate generating module configured to collect and analyze the behavior data to be detected generated on the device screen, and obtain the horizontal coordinate list and the vertical coordinate list left by the behavior data to be detected on the device screen;
  • a trajectory generation module configured to obtain a trajectory sequence corresponding to the behavior data to be detected according to the horizontal coordinate list, the vertical coordinate list, and an initialization trajectory sequence;
  • a quantity calculation module configured to store the trajectory sequence in a buffer, and perform a summation and count on the trajectory sequence in real time;
  • the trajectory determination module is configured to compare the count value with a preset threshold value, and determine a trajectory sequence whose count value exceeds the preset threshold value as a scalper trajectory.
  • the device further includes:
  • An image division module configured to equally divide the screen image of the device screen for which the behavior data to be detected is to be collected into m rows and n columns of grid-like images, where m and n are positive integers;
  • the sequence generation module is used to generate an initialization sequence according to the grid image, wherein the initialization trajectory sequence is m*n characters '0'.
  • the device further includes:
  • the resolution verification module is used to obtain the resolution of the device screen and verify whether the resolution is consistent with the preset resolution
  • the resolution conversion module is used to convert the resolution of the screen image into a preset resolution.
  • trajectory generation module includes:
  • a position calculation unit configured to traverse the horizontal coordinate list and the vertical coordinate list, and calculate the corresponding position of each coordinate in the initialization trajectory sequence
  • the character filling unit is used to fill the character "1" in the corresponding position to generate the trajectory sequence corresponding to the behavior data to be detected.
  • the device further includes:
  • the trajectory comparison module is used to compare the acquired new trajectory sequence with the scalper trajectory, and if they are consistent, determine the new trajectory sequence as the scalper trajectory.
  • the method and device for detecting similar human-machine trajectories based on screen segmentation obtain the horizontal coordinate list and the vertical coordinate list left by the behavior data to be detected on the device screen by parsing the behavior data to be detected , Combine the horizontal coordinate list, the vertical coordinate list and the initialization sequence to obtain the user's behavior trajectory sequence, and then count the trajectory sequence, set the threshold, and detect whether the trajectory sequence is a scalper trajectory, which effectively prevents the scalper’s scalping behavior and reduces the scalper Economic losses caused;
  • the human-machine similar trajectory detection method and device based on screen segmentation provided by the embodiments of the present invention obtain the initialization sequence by segmenting the device screen, and is used to subsequently obtain the trajectory sequence corresponding to the behavior data to be detected, reducing the local coordinate system The impact of the up, down, left, and right offsets on the detection and improve the detection accuracy;
  • the human-machine similar trajectory detection method and device based on screen segmentation provided by the embodiments of the present invention unify the device screen resolution to the same resolution through calculation, and shield the difference in resolution of different devices;
  • the human-machine similar trajectory detection method and device based on screen segmentation provided by the embodiments of the present invention subsequently compare the acquired user-generated trajectory sequence with the scalper trajectory sequence to determine whether the new trajectory sequence is a scalper trajectory , On the one hand to improve the efficiency of detection, on the other hand to improve the reusability of data.
  • Fig. 1 is a flowchart showing a method for detecting similar human-machine trajectories based on screen segmentation according to an exemplary embodiment
  • Fig. 2 is a schematic structural diagram of a human-machine similar trajectory detection device based on screen segmentation according to an exemplary embodiment.
  • Fig. 1 is a flow chart showing a method for detecting human-machine similar trajectories based on screen segmentation according to an exemplary embodiment. Referring to Fig. 1, the method includes the following steps:
  • S1 Collect and analyze the behavior data to be detected generated on the device screen, and obtain the horizontal coordinate list and the vertical coordinate list left by the behavior data to be detected on the device screen.
  • the e-commerce platform interface etc.
  • the coordinates include the abscissa and the corresponding ordinate, and each set of coordinates represents a point on the device screen. All the abscissas and ordinates will be grouped together to generate the corresponding list of horizontal and vertical coordinates. For specific calculations, you can establish a two-dimensional coordinate system on the device screen, and then use the two-dimensional coordinate system as the basis to calculate the coordinates of the behavior data to be detected on the screen, that is, the trajectory behavior left by the user on the device screen. coordinate.
  • the behavior data to be detected is converted into a trajectory sequence, and then each trajectory sequence is mutually performed.
  • each trajectory sequence is mutually performed.
  • an initialization trajectory sequence is set first, and then the trajectory sequence corresponding to the behavior data to be detected is calculated according to the horizontal coordinate list, the vertical coordinate list and the initialization trajectory sequence obtained in the above steps.
  • S3 Store the trajectory sequence in a buffer, and perform a sum and count on the trajectory sequence in real time.
  • each acquired trajectory sequence is stored in the buffer, and then each trajectory sequence is summed and counted in real time.
  • the specific method of summing and counting includes: comparing the currently acquired trajectory sequence with each trajectory sequence in the buffer, and every time a trajectory sequence is found to be the same as the current trajectory sequence, the count of the current trajectory sequence is increased by one.
  • S4 The count value is compared with a preset threshold value, and the track sequence in which the count value exceeds the preset threshold value is determined as a scalper track.
  • an appropriate preset threshold is set, and the count value of the trajectory sequence is compared with the preset threshold.
  • the trajectory sequence is judged as a scalper trajectory .
  • the basis for this determination is that in reality, the probability that a large number of trajectory behaviors left by the user on the device screen is the same is very low.
  • an appropriate preset threshold can be set, and when the count value of a certain trajectory sequence exceeds the preset threshold, the trajectory sequence is judged as a scalper trajectory (ie, a false trajectory).
  • the specific value of the preset threshold is not limited, and the user can set it according to actual needs, and the preset threshold in the embodiment of the present invention supports dynamic adjustment.
  • the method further includes:
  • the screen image of the device screen for which the behavior data to be detected is to be collected is equally divided into m rows and n columns of grid-like images, where m and n are positive integers.
  • the screen image of the device screen that generates the behavior data to be detected is obtained, and the screen image is equally divided into n columns in the horizontal direction and m rows in the vertical direction, that is, the screen image is divided into m*n grid-like images of small grids, where Both m and n are positive integers, and the values of m and n are affected by the width ⁇ x and height ⁇ y of the above-mentioned small cells.
  • the greater the value of height ⁇ y, m The smaller the value.
  • the value of m is 10 and the value of n is 6, that is, the screen image is equally divided into 60 parts. It should be noted here that in the embodiment of the present invention, the values of m and n are not limited, and the user can set them according to actual needs.
  • An initialization sequence is generated according to the grid-like image, wherein the initialization trajectory sequence is m*n characters '0'.
  • an initialization trajectory sequence is set first.
  • the initialization sequence is generated according to the grid-like image, and the initialization trajectory sequence is set as a string of character strings "0000", in which the number of characters '0' is m*n, that is, the number of characters in the initialization and The number of small grids in the grid-like image is the same, and each character '0' corresponds to a grid in the grid-like image.
  • the method before dividing the screen of the device whose behavior data to be detected is to be collected into a grid-like image of m rows and n columns, the method further includes:
  • the collected user-generated behavior data to be detected may come from different devices, and the screen resolution of different devices may be different, and the difference in resolution will cause differences in the acquired trajectory sequence, resulting in detection of The accuracy is reduced. Therefore, in order to shield the difference in the resolution of different devices, in the embodiment of the present invention, the resolution of the screen image of the device screen that generates the behavior data to be detected is unified to the same resolution, and then the device screen is divided It is a grid-like image.
  • obtaining the trajectory sequence corresponding to the behavior data to be detected includes:
  • the horizontal coordinate list and the vertical coordinate list obtained in the above steps are traversed, and the position of the small grid in the grid-like image corresponding to each coordinate is calculated. Since each character '0' in the initialization sequence corresponds to the grid-like image Therefore, the corresponding position of each coordinate in the initialization trajectory sequence can be calculated, and then the character "0" is changed to the character "1" in the position corresponding to the initialization trajectory sequence to generate the trajectory corresponding to the behavior data to be detected sequence. Since the user’s trajectory behavior is converted into a trajectory sequence, and each coordinate is calculated corresponding to the position of the small grid in the grid-like image, rather than the position of the point, it can reduce the offset of the local coordinates up, down, left, and right. The impact on the test results, and improve the accuracy of the test.
  • the method further includes:
  • the obtained new trajectory sequence is compared with the scalper trajectory, and if they are consistent, the new trajectory sequence is determined as the scalper trajectory.
  • it may also be configured to compare the acquired new trajectory sequence with the scalper trajectory. If they are consistent, the new trajectory sequence is considered to be problematic, and the new trajectory sequence is determined as the scalper trajectory. By comparing the acquired user-generated trajectory sequence with the scalper trajectory sequence, it is determined whether the new trajectory sequence is a scalper trajectory. On the one hand, the detection efficiency can be improved, and the data reusability can be improved on the other hand.
  • Fig. 2 is a schematic diagram showing the structure of a human-machine similar trajectory detection device based on screen segmentation according to an exemplary embodiment. As described with reference to Fig. 2, the device includes:
  • a coordinate generating module configured to collect and analyze the behavior data to be detected generated on the device screen, and obtain the horizontal coordinate list and the vertical coordinate list left by the behavior data to be detected on the device screen;
  • a trajectory generation module configured to obtain a trajectory sequence corresponding to the behavior data to be detected according to the horizontal coordinate list, the vertical coordinate list, and an initialization trajectory sequence;
  • a quantity calculation module configured to store the trajectory sequence in a buffer, and perform a summation and count on the trajectory sequence in real time;
  • the trajectory determination module is configured to compare the count value with a preset threshold value, and determine a trajectory sequence whose count value exceeds the preset threshold value as a scalper trajectory.
  • the device further includes:
  • An image division module configured to equally divide the screen image of the device screen for which the behavior data to be detected is to be collected into m rows and n columns of grid-like images, where m and n are positive integers;
  • the sequence generation module is used to generate an initialization sequence according to the grid image, wherein the initialization trajectory sequence is m*n characters '0'.
  • the device further includes:
  • the resolution verification module is used to obtain the resolution of the device screen and verify whether the resolution is consistent with the preset resolution
  • the resolution conversion module is used to convert the resolution of the screen image into a preset resolution.
  • the trajectory generation module includes:
  • a position calculation unit configured to traverse the horizontal coordinate list and the vertical coordinate list, and calculate the corresponding position of each coordinate in the initialization trajectory sequence
  • the character filling unit is used to fill the character "1" in the corresponding position to generate the trajectory sequence corresponding to the behavior data to be detected.
  • the device further includes:
  • the trajectory comparison module is used to compare the acquired new trajectory sequence with the scalper trajectory, and if they are consistent, determine the new trajectory sequence as the scalper trajectory.
  • the method and device for detecting similar human-machine trajectories based on screen segmentation obtain the horizontal coordinate list and the vertical coordinate list left by the behavior data to be detected on the device screen by parsing the behavior data to be detected , Combine the horizontal coordinate list, the vertical coordinate list and the initialization sequence to obtain the user's behavior trajectory sequence, and then count the trajectory sequence, set the threshold, and detect whether the trajectory sequence is a scalper trajectory, which effectively prevents the scalper’s scalping behavior and reduces the scalper Economic losses caused;
  • the human-machine similar trajectory detection method and device based on screen segmentation provided by the embodiments of the present invention obtain the initialization sequence by segmenting the device screen, and is used to subsequently obtain the trajectory sequence corresponding to the behavior data to be detected, reducing the local coordinate system The impact of the up, down, left, and right offsets on the detection and improve the detection accuracy;
  • the human-machine similar trajectory detection method and device based on screen segmentation provided by the embodiments of the present invention unify the device screen resolution to the same resolution through calculation, and shield the difference in resolution of different devices;
  • the human-machine similar trajectory detection method and device based on screen segmentation provided by the embodiments of the present invention subsequently compare the acquired user-generated trajectory sequence with the scalper trajectory sequence to determine whether the new trajectory sequence is a scalper trajectory , On the one hand to improve the efficiency of detection, on the other hand to improve the reusability of data.
  • the human-machine similar trajectory detection device based on screen segmentation only uses the division of the above functional modules to illustrate when the detection service is triggered.
  • the above functions can be allocated according to needs. It is completed by different functional modules, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.
  • the device for detecting similar human-machine trajectories based on screen segmentation provided by the above embodiments belong to the same concept as the embodiment of the method for detecting similar human-machine trajectories based on screen segmentation, that is, the device is based on the method for detecting similar human-machine trajectories based on screen segmentation. Yes, for the specific implementation process, please refer to the method embodiment, which will not be repeated here.
  • the program can be stored in a computer-readable storage medium.
  • the storage medium mentioned can be a read-only memory, a magnetic disk or an optical disk, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Business, Economics & Management (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Image Analysis (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

一种基于屏幕分割的人机相似轨迹检测方法及装置，该方法包括：采集并解析设备屏幕上产生的待检测行为数据，获取待检测行为数据在设备屏幕上留下的横向坐标列表以及纵向坐标列表（S1）；根据横向坐标列表、纵向坐标列表以及初始化轨迹序列，获取待检测行为数据所对应的轨迹序列（S2）；将轨迹序列存储至缓存中，并实时对轨迹序列进行求和计数（S3）；将计数的值与预设阈值进行比较，将计数的值超过预设阈值的轨迹序列判定为黄牛轨迹（S4）。上述方法通过基于屏幕设备分辨率进行网格划分，并使用一串轨迹序列表示用户的轨迹行为，提高检测的精度，并有效制止了黄牛的刷单行为，降低了黄牛带来的经济损失。

Description

一种基于屏幕分割的人机相似轨迹检测方法及装置 技术领域
本发明涉及电商风控技术领域，特别涉及一种基于屏幕分割的人机相似轨迹检测方法及装置。
背景技术
在电商环境里，黄牛通过脚本刷单与抢券的现象大量存在，给商家与平台都带来严重的损失。目前对相似轨迹识别的手段是将采集的人机数据转化为MD5值，然后通过比较MD5值来判断若干条轨迹是否相似。MD5，即Message DigestAlgorithm MD5，中文名为消息摘要算法第五版，是计算机安全领域广泛使用的一种散列函数，用以提供消息的完整性保护。MD5用于确保信息传输完整一致，是计算机广泛使用的杂凑算法之一(又译摘要算法、哈希算法)，主要用于将数据(如汉字)运算为另一固定长度值。
但是，上述方法主要缺陷是只能发现完全一样的人机轨迹，而目前的人机轨迹特点是，回放的轨迹图像整体相似，而局部坐标具有上下左右的偏移量，很难有两条完全一模一样的轨迹。因此，检测的精确度不高。
发明内容
为了解决现有技术的问题，本发明实施例提供了一种基于屏幕分割的人机相似轨迹检测方法，以克服现有技术中回放的轨迹图像整体相似，而局部坐标具有上下左右的偏移量，很难有两条完全一模一样的轨迹，导致检测的精确度不高等问题。
为解决上述一个或多个技术问题，本发明采用的技术方案是：
一方面，提供了一种基于屏幕分割的人机相似轨迹检测方法，该方法包括如下步骤：
采集并解析所述设备屏幕上产生的待检测行为数据，获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表；
根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列；
将所述轨迹序列存储至缓存中，并实时对所述轨迹序列进行求和计数；
将所述计数的值与预设阈值进行比较，将计数的值超过所述预设阈值的轨迹序列判定为黄牛轨迹。
进一步的，所述方法还包括：
将待采集待所述检测行为数据的设备屏幕的屏幕图像等分成m行n列个小格的网格状图像，其中m和n为正整数；
根据所述网格状图像生成初始化序列，其中，所述初始化轨迹序列为m*n个字符‘0’。
进一步的，将待采集待检测行为数据的设备屏幕等分成m行n列个小格的网格状图像前，所述方法还包括：
获取所述设备屏幕的分辨率，验证所述分辨率是否与预设的分辨率一致，若不一致，则将所述屏幕图像的分辨率转换为预设的分辨率。
进一步的，根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列包括：
遍历所述横向坐标列表以及所述纵向坐标列表，计算每个坐标在所述初始化轨迹序列中对应的位置，在所述对应的位置中填充字符“1”，生成所述待检测行为数据所对应的轨迹序列。
进一步的，所述方法还包括：
将获取到的新的轨迹序列与所述黄牛轨迹进行比较，若一致，则将所述新的轨迹序列判定为黄牛轨迹。
另一方面，提供了一种基于屏幕分割的人机相似轨迹检测装置，所述装置包括：
坐标生成模块，用于采集并解析所述设备屏幕上产生的待检测行为数据，获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表；
轨迹生成模块，用于根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列；
数量计算模块，用于将所述轨迹序列存储至缓存中，并实时对所述轨迹序列进行求和计数；
轨迹判定模块，用于将所述计数的值与预设阈值进行比较，将计数的值超过所述预设阈值的轨迹序列判定为黄牛轨迹。
进一步的，所述装置还包括：
图像划分模块，用于将待采集所述待检测行为数据的设备屏幕的屏幕图像等分成m行n列个小格的网格状图像，其中m和n为正整数；
序列生成模块，用于根据所述网格状图像生成初始化序列，其中，所述初始化轨迹序列为m*n个字符‘0’。
进一步的，所述装置还包括：
分辨率验证模块，用于获取所述设备屏幕的分辨率，验证所述分辨率是否与预设的分辨率一致；
分辨率转换模块，用于将所述屏幕图像的分辨率转换为预设的分辨率。
进一步的，所述轨迹生成模块包括：
位置计算单元，用于遍历所述横向坐标列表以及所述纵向坐标列表，计算每个坐标在所述初始化轨迹序列中对应的位置；
字符填充单元，用于在所述对应的位置中填充字符“1”，生成所述待检测行为数据所对应的轨迹序列。
进一步的，所述装置还包括：
轨迹比较模块，用于将获取到的新的轨迹序列与所述黄牛轨迹进行比较，若一致，则将所述新的轨迹序列判定为黄牛轨迹。
本发明实施例提供的技术方案带来的有益效果是：
1、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，通过解析待检测行为数据获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表，结合横向坐标列表以及纵向坐标列表以及初始化序列获得用户的行为轨迹序列，然后对轨迹序列计数，设置阀值，检测出轨迹序列是否为黄牛轨迹，有效制止了黄牛的刷单行为，降低了黄牛带来的经济损失；
2、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，通过对设备屏幕进行分割的方式获取初始化序列，并用于后续获取待检测行为数据对应的轨迹序列，降低局部坐标具有上下左右的偏移量给检测带来的影响，提高检测的精度；
3、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，通过计算将设备屏幕分辨率统一到同一个分辨率上，屏蔽不同设备的分辨率的差异性；
4、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，后续通过将获取到的用户生成的轨迹序列与黄牛轨迹序列进行比较，从而判断该新的轨迹序列是否为黄牛轨迹，一方面提高检测的效率，另一方面提高数据的复用性。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是根据一示例性实施例示出的基于屏幕分割的人机相似轨迹检测方法的流程图；
图2是根据一示例性实施例示出的基于屏幕分割的人机相似轨迹检测装置的结构示意图。
具体实施方式
为使本发明的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
图1是根据一示例性实施例示出的基于屏幕分割的人机相似轨迹检测方法的流程图，参照图1所示，该方法包括如下步骤：
S1：采集并解析所述设备屏幕上产生的待检测行为数据，获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表。
具体的，通过调用电商平台接口等方式，获取用户在设备屏幕上留下的待检测行为数据，对该待检测行为数据进行解析，获取待检测行为数据在设备屏幕上留下所有的坐标，该坐标包括横坐标和与之对应的纵坐标，每组坐标都代表设备屏幕上的一个点。将分别将所有的横坐标和纵坐标集合在一起，生成相应的横向坐标列表以及纵向坐标列表。具体计算时，可以在设备屏幕上建立二维坐标系，然后以该二维坐标系为依据，计算待检测行为数据在屏幕上留下的坐标，即用户在设备屏幕上留下的轨迹行为的坐标。
S2：根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列。
具体的，为了后续方便对采集到的用户留在设备屏幕上的轨迹行为进行比对，本发明实施例中，采用将待检测行为数据转换为轨迹序列，再将每条轨迹 序列之间进行相互比对的方式。具体实施的时候，会先设置一个初始化轨迹序列，然后根据上述步骤获取到的横向坐标列表、纵向坐标列表以及初始化轨迹序列计算获取待检测行为数据所对应的轨迹序列。
S3：将所述轨迹序列存储至缓存中，并实时对所述轨迹序列进行求和计数。
具体的，将获取到的每条轨迹序列均存储至缓存中，然后实时对每条轨迹序列进行求和计数。具体求和计数的方式包括：将当前获取到的轨迹序列与缓存中的每一条轨迹序列进行比较，每发现一条轨迹序列与当前轨迹序列相同，则将当前轨迹序列的计数加1。
S4：将所述计数的值与预设阈值进行比较，将计数的值超过所述预设阈值的轨迹序列判定为黄牛轨迹。
具体的，设置一个合适的预设阈值，将轨迹序列的计数的值与预设阈值进行比较，当某条轨迹序列的计数的值超过该预设阈值，则将该条轨迹序列判定为黄牛轨迹。这样判定的依据是，现实中用户在设备屏幕上留下的轨迹行为存在大量相同的概率很低，当获取到的大量的轨迹行为相同时，就说明这些轨迹行为很可疑，可能是黄牛通过脚本实现的，因此，可以设置一个合适的预设阈值，当某条轨迹序列的计数的值超过该预设阈值，则将该条轨迹序列判定为黄牛轨迹(即虚假轨迹)。这里需要说明的是，本发明实施例中，对预设阈值的具体数值不进行限制，用户可以根据实际需求进行设置，并且本发明实施例中的预设阈值支持动态调整。
作为一种较优的实施方式，本发明实施例中，所述方法还包括：
将待采集所述待检测行为数据的设备屏幕的屏幕图像等分成m行n列个小格的网格状图像，其中m和n为正整数。
具体的，获取产生待检测行为数据的设备屏幕的屏幕图像，将该屏幕图像横向等分成n列，纵向等分成m行，即将屏幕图像划分为m*n个小格的网格状图像，其中m和n均为正整数，且m和n的值受上述小格的宽Δx与高Δy的影响，宽Δx的值越大，n的值越小，同样，高Δy的值越大，m的值越小。
作为一种较优的实施方式，m取值为10，n取值为6，即将屏幕图像等分成60份。这里需要说明的是，本发明实施例中，对m和n取值不做限制，用户可以根据实际需求进行设置。
根据所述网格状图像生成初始化序列，其中，所述初始化轨迹序列为m*n个字符‘0’。
具体的，本发明实施例中，使用一串轨迹序列表示用户的轨迹行为，因此，先设置一初始化轨迹序列。具体实施时，根据网格状图像生成初始化序列，设置初始化轨迹序列为一串字符串“0000……”，其中字符‘0’的个数为m*n个，即初始化中字符的个数与网格状图像中的小格数量一致，每一个字符‘0’均对应于网格状图像中的一格。
作为一种较优的实施方式，本发明实施例中，将待采集待检测行为数据的设备屏幕等分成m行n列个小格的网格状图像前，所述方法还包括：
获取所述设备屏幕的分辨率，验证所述分辨率是否与预设的分辨率一致，若不一致，则将所述屏幕图像的分辨率转换为预设的分辨率。
具体的，采集到的用户产生的的待检测行为数据可能来自不同的设备，不同的设备屏幕的分辨率可能有所不同，而分辨率的不同会造成获取到的轨迹序列有差异，导致检测的精度下降。因此，为了屏蔽不同设备的分辨率的差异性，本发明实施例中，采用先将产生待检测行为数据的设备屏幕的屏幕图像的分辨率统一到同一个分辨率上，然后在将设备屏幕划分为网格状图像。
作为一种较优的实施方式，本发明实施例中，根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列包括：
遍历所述横向坐标列表以及所述纵向坐标列表，计算每个坐标在所述初始化轨迹序列中对应的位置，在所述对应的位置中填充字符“1”，生成所述待检测行为数据所对应的轨迹序列。
具体的，遍历上述步骤获取到的横向坐标列表以及纵向坐标列表，计算每 个坐标对应的网格状图像中小格的位置，由于初始化序列中每一个字符‘0’均对应于网格状图像中的一格，因此可以计算出每个坐标在初始化轨迹序列中对应的位置，然后在初始化轨迹序列对应的位置中将字符“0”修改为字符“1”，生成待检测行为数据所对应的轨迹序列。由于将用户的轨迹行为转换为了轨迹序列，且计算的是每个坐标对应于网格状图像中的小格的位置，而不是点的位置，因此，可以降低局部坐标具有上下左右的偏移量给检测结果带来的影响，提高检测的精度。
作为一种较优的实施方式，本发明实施例中，所述方法还包括：
将获取到的新的轨迹序列与所述黄牛轨迹进行比较，若一致，则将所述新的轨迹序列判定为黄牛轨迹。
具体的，本发明实施例中，还可以设置将获取到的新的轨迹序列与黄牛轨迹进行比较，若一致，则认为该条新的轨迹序列有问题，将新的轨迹序列判定为黄牛轨迹。通过将获取到的用户生成的轨迹序列与黄牛轨迹序列进行比较，从而判断该新的轨迹序列是否为黄牛轨迹，一方面提高检测的效率，另一方面可以提高数据的复用性。
图2是根据一示例性实施例示出的基于屏幕分割的人机相似轨迹检测装置的结构示意图，参照图2所述，该装置包括：
坐标生成模块，用于采集并解析所述设备屏幕上产生的待检测行为数据，获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表；
轨迹生成模块，用于根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列；
数量计算模块，用于将所述轨迹序列存储至缓存中，并实时对所述轨迹序列进行求和计数；
轨迹判定模块，用于将所述计数的值与预设阈值进行比较，将计数的值超过所述预设阈值的轨迹序列判定为黄牛轨迹。
作为一种较优的实施方式，本发明实施例中，所述装置还包括：
图像划分模块，用于将待采集所述待检测行为数据的设备屏幕的屏幕图像等分成m行n列个小格的网格状图像，其中m和n为正整数；
序列生成模块，用于根据所述网格状图像生成初始化序列，其中，所述初始化轨迹序列为m*n个字符‘0’。
作为一种较优的实施方式，本发明实施例中，所述装置还包括：
分辨率验证模块，用于获取所述设备屏幕的分辨率，验证所述分辨率是否与预设的分辨率一致；
分辨率转换模块，用于将所述屏幕图像的分辨率转换为预设的分辨率。
作为一种较优的实施方式，本发明实施例中，所述轨迹生成模块包括：
位置计算单元，用于遍历所述横向坐标列表以及所述纵向坐标列表，计算每个坐标在所述初始化轨迹序列中对应的位置；
字符填充单元，用于在所述对应的位置中填充字符“1”，生成所述待检测行为数据所对应的轨迹序列。
作为一种较优的实施方式，本发明实施例中，所述装置还包括：
轨迹比较模块，用于将获取到的新的轨迹序列与所述黄牛轨迹进行比较，若一致，则将所述新的轨迹序列判定为黄牛轨迹。
综上所述，本发明实施例提供的技术方案带来的有益效果是：
1、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，通过解析待检测行为数据获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表，结合横向坐标列表以及纵向坐标列表以及初始化序列获得用户的行为轨迹序列，然后对轨迹序列计数，设置阀值，检测出轨迹序列是否为黄牛轨迹，有效制止了黄牛的刷单行为，降低了黄牛带来的经济损失；
2、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，通过对设备屏幕进行分割的方式获取初始化序列，并用于后续获取待检测行为数 据对应的轨迹序列，降低局部坐标具有上下左右的偏移量给检测带来的影响，提高检测的精度；
3、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，通过计算将设备屏幕分辨率统一到同一个分辨率上，屏蔽不同设备的分辨率的差异性；
4、本发明实施例提供的基于屏幕分割的人机相似轨迹检测方法及装置，后续通过将获取到的用户生成的轨迹序列与黄牛轨迹序列进行比较，从而判断该新的轨迹序列是否为黄牛轨迹，一方面提高检测的效率，另一方面提高数据的复用性。
需要说明的是：上述实施例提供的基于屏幕分割的人机相似轨迹检测装置在触发检测业务时，仅以上述各功能模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能模块完成，即将装置的内部结构划分成不同的功能模块，以完成以上描述的全部或者部分功能。另外，上述实施例提供的基于屏幕分割的人机相似轨迹检测装置与基于屏幕分割的人机相似轨迹检测方法实施例属于同一构思，即该装置是基于该基于屏幕分割的人机相似轨迹检测方法的，其具体实现过程详见方法实施例，这里不再赘述。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。
以上所述仅为本发明的较佳实施例，并不用以限制本发明，凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (10)

  1. 一种基于屏幕分割的人机相似轨迹检测方法，其特征在于，所述方法包括如下步骤：
    采集并解析所述设备屏幕上产生的待检测行为数据，获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表；
    根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列；
    将所述轨迹序列存储至缓存中，并实时对所述轨迹序列进行求和计数；
    将所述计数的值与预设阈值进行比较，将计数的值超过所述预设阈值的轨迹序列判定为黄牛轨迹。
  2. 根据权利要求1所述的基于屏幕分割的人机相似轨迹检测方法，其特征在于，所述方法还包括：
    将待采集所述待检测行为数据的设备屏幕的屏幕图像等分成m行n列个小格的网格状图像，其中m和n为正整数；
    根据所述网格状图像生成初始化序列，其中，所述初始化轨迹序列为m*n个字符‘0’。
  3. 根据权利要求2所述的基于屏幕分割的人机相似轨迹检测方法，其特征在于，将待采集待检测行为数据的设备屏幕等分成m行n列个小格的网格状图像前，所述方法还包括：
    获取所述设备屏幕的分辨率，验证所述分辨率是否与预设的分辨率一致，若不一致，则将所述屏幕图像的分辨率转换为预设的分辨率。
  4. 根据权利要求1至3任一所述的基于屏幕分割的人机相似轨迹检测方法，其特征在于，根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列包括：
    遍历所述横向坐标列表以及所述纵向坐标列表，计算每个坐标在所述初始 化轨迹序列中对应的位置，在所述对应的位置中填充字符“1”，生成所述待检测行为数据所对应的轨迹序列。
  5. 根据权利要求1至3任一所述的基于屏幕分割的人机相似轨迹检测方法，其特征在于，所述方法还包括：
    将获取到的新的轨迹序列与所述黄牛轨迹进行比较，若一致，则将所述新的轨迹序列判定为黄牛轨迹。
  6. 一种基于屏幕分割的人机相似轨迹检测装置，其特征在于，所述装置包括：
    坐标生成模块，用于采集并解析所述设备屏幕上产生的待检测行为数据，获取所述待检测行为数据在所述设备屏幕上留下的横向坐标列表以及纵向坐标列表；
    轨迹生成模块，用于根据所述横向坐标列表、所述纵向坐标列表以及初始化轨迹序列，获取所述待检测行为数据所对应的轨迹序列；
    数量计算模块，用于将所述轨迹序列存储至缓存中，并实时对所述轨迹序列进行求和计数；
    轨迹判定模块，用于将所述计数的值与预设阈值进行比较，将计数的值超过所述预设阈值的轨迹序列判定为黄牛轨迹。
  7. 根据权利要求6所述的基于屏幕分割的人机相似轨迹检测装置，其特征在于，所述装置还包括：
    图像划分模块，用于将待采集所述待检测行为数据的设备屏幕的屏幕图像等分成m行n列个小格的网格状图像，其中m和n为正整数；
    序列生成模块，用于根据所述网格状图像生成初始化序列，其中，所述初始化轨迹序列为m*n个字符‘0’。
  8. 根据权利要求7所述的基于屏幕分割的人机相似轨迹检测装置，其特征在于，所述装置还包括：
    分辨率验证模块，用于获取所述设备屏幕的分辨率，验证所述分辨率是否 与预设的分辨率一致；
    分辨率转换模块，用于将所述屏幕图像的分辨率转换为预设的分辨率。
  9. 根据权利要求6或7所述的基于屏幕分割的人机相似轨迹检测装置，其特征在于，所述轨迹生成模块包括：
    位置计算单元，用于遍历所述横向坐标列表以及所述纵向坐标列表，计算每个坐标在所述初始化轨迹序列中对应的位置；
    字符填充单元，用于在所述对应的位置中填充字符“1”，生成所述待检测行为数据所对应的轨迹序列。
  10. 根据权利要求6或7所述的基于屏幕分割的人机相似轨迹检测装置，其特征在于，所述装置还包括：
    轨迹比较模块，用于将获取到的新的轨迹序列与所述黄牛轨迹进行比较，若一致，则将所述新的轨迹序列判定为黄牛轨迹。
PCT/CN2020/097854 2019-08-29 2020-06-24 一种基于屏幕分割的人机相似轨迹检测方法及装置 WO2021036454A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CA3152854A CA3152854C (en) 2019-08-29 2020-06-24 Human-computer similar trajectory detection method and apparatus based on screen segmentation

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910811578.8A CN110689400B (zh) 2019-08-29 2019-08-29 一种基于屏幕分割的人机相似轨迹检测方法及装置
CN201910811578.8 2019-08-29

Publications (1)

Publication Number Publication Date
WO2021036454A1 true WO2021036454A1 (zh) 2021-03-04

Family

ID=69107650

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/097854 WO2021036454A1 (zh) 2019-08-29 2020-06-24 一种基于屏幕分割的人机相似轨迹检测方法及装置

Country Status (3)

Country Link
CN (1) CN110689400B (zh)
CA (1) CA3152854C (zh)
WO (1) WO2021036454A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115062695A (zh) * 2022-06-06 2022-09-16 北京字跳网络技术有限公司 一种作弊判断方法、装置、设备及介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110689400B (zh) * 2019-08-29 2022-02-25 苏宁云计算有限公司 一种基于屏幕分割的人机相似轨迹检测方法及装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090184809A1 (en) * 2008-01-09 2009-07-23 Cybernet Systems Corporation Rapid prototyping and machine vision for reconfigurable interfaces
US20170228833A1 (en) * 2016-02-08 2017-08-10 Jonathan Previtali Solar Services Matchmaking System
CN107679374A (zh) * 2017-08-23 2018-02-09 北京三快在线科技有限公司 一种基于滑动轨迹的人机识别方法及装置，电子设备
CN109388934A (zh) * 2018-09-10 2019-02-26 平安科技（深圳）有限公司 信息验证方法、装置、计算机设备及存储介质
CN110119602A (zh) * 2018-02-06 2019-08-13 百度在线网络技术（北京）有限公司 人机识别方法、装置、服务器、客户端及存储介质
CN110689400A (zh) * 2019-08-29 2020-01-14 苏宁云计算有限公司 一种基于屏幕分割的人机相似轨迹检测方法及装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107341384A (zh) * 2016-04-28 2017-11-10 苏宁云商集团股份有限公司 一种业务验证方法及系统
CN106529269A (zh) * 2016-12-01 2017-03-22 微梦创科网络科技（中国）有限公司 一种安全验证方法及系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090184809A1 (en) * 2008-01-09 2009-07-23 Cybernet Systems Corporation Rapid prototyping and machine vision for reconfigurable interfaces
US20170228833A1 (en) * 2016-02-08 2017-08-10 Jonathan Previtali Solar Services Matchmaking System
CN107679374A (zh) * 2017-08-23 2018-02-09 北京三快在线科技有限公司 一种基于滑动轨迹的人机识别方法及装置，电子设备
CN110119602A (zh) * 2018-02-06 2019-08-13 百度在线网络技术（北京）有限公司 人机识别方法、装置、服务器、客户端及存储介质
CN109388934A (zh) * 2018-09-10 2019-02-26 平安科技（深圳）有限公司 信息验证方法、装置、计算机设备及存储介质
CN110689400A (zh) * 2019-08-29 2020-01-14 苏宁云计算有限公司 一种基于屏幕分割的人机相似轨迹检测方法及装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115062695A (zh) * 2022-06-06 2022-09-16 北京字跳网络技术有限公司 一种作弊判断方法、装置、设备及介质

Also Published As

Publication number Publication date
CN110689400B (zh) 2022-02-25
CA3152854A1 (en) 2021-03-04
CN110689400A (zh) 2020-01-14
CA3152854C (en) 2024-01-02

Similar Documents

Publication Publication Date Title
US10592667B1 (en) Methods and apparatus for detecting malware samples with similar image sets
TWI515598B (zh) 產生純化惡意程式的方法、偵測惡意程式之方法及其系統
WO2021036454A1 (zh) 一种基于屏幕分割的人机相似轨迹检测方法及装置
US9122341B2 (en) Resolving merged touch contacts
Kang et al. Model assertions for debugging machine learning
US6523141B1 (en) Method and apparatus for post-mortem kernel memory leak detection
US9251352B2 (en) Quantitative analysis of information leakage vulnerabilities
CN112866023A (zh) 网络检测、模型训练方法、装置、设备及存储介质
US20070247937A1 (en) Information processing system for calculating the number of redundant lines optimal for memory device
US11531748B2 (en) Method and system for autonomous malware analysis
KR20200080541A (ko) 프로그램 경로에 기반한 소프트웨어 취약점 검출 장치 및 방법
CN112306833A (zh) 应用程序的崩溃统计方法、装置、计算机设备及存储介质
CN107920067B (zh) 一种主动对象存储系统上的入侵检测方法
Qutub et al. Hardware faults that matter: Understanding and estimating the safety impact of hardware faults on object detection dnns
CN115146275A (zh) 容器安全防护方法、装置、电子设备和存储介质
CN105302700A (zh) 一种记录在触摸终端上的用户操作的方法与设备
CN113625924A (zh) 基于移动操作的元素选取方法、设备、存储介质及装置
WO2022170170A1 (en) Malware detection system
Weigel et al. Kernel vulnerability factor and efficient hardening for histogram of oriented gradients
CN114328171A (zh) 测试用例库的生成方法、装置、设备及存储介质
EP3671467A1 (en) Gui application testing using bots
US20240096079A1 (en) Method for detecting image sizes, computer device, and storage medium
US11822655B1 (en) False alarm reduction by novelty detection
CN116488943B (zh) 多媒体数据泄露溯源检测方法、装置及设备
Tian et al. A malicious application detection model to remove the influence of interference API sequence

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: 20859402

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 3152854

Country of ref document: CA

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20859402

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20859402

Country of ref document: EP

Kind code of ref document: A1