WO2020125370A1

WO2020125370A1 - Relationship analysis method and related product

Info

Publication number: WO2020125370A1
Application number: PCT/CN2019/121610
Authority: WO
Inventors: 岑助甫; 郑承渊; 黄永熙; 凌睿; 母凤轩; 罗同舟
Original assignee: 深圳云天励飞技术有限公司
Priority date: 2018-12-21
Filing date: 2019-11-28
Publication date: 2020-06-25
Also published as: CN109766786A; CN109766786B

Abstract

A relationship analysis method and a related product. The method comprises: acquiring a companionship parameter with respect to a target user and at least one companion user of the target user (201); establishing, according to the companionship parameter of the at least one companion user, a target relationship graph with respect to the target user and the at least one companion user (202); and determining, according to the target relationship graph, a target user relationship between the target user and the at least one companion user (203). Thus, the invention intuitively reflects relationships between users, and improves the efficiency of determining user relationships.

Description

Character relationship analysis method and related products

This application requires priority to be submitted to the Chinese Patent Office on December 21, 2018, with the application number 201811573297.5 and the invention titled "Person Relation Analysis Method and Related Products". The entire content of this application is incorporated by reference in this application .

Technical field

This application relates to the field of data processing technology, and specifically relates to a character relationship analysis method and related products.

Background technique

With the continuous development of urban civilization, the relationship between people has become more and more complicated. At present, when analyzing the relationship between users, the main method is to obtain the existing data, and then analyze the existing data to obtain the relationship between the users. When obtaining the relationship between the users It is inefficient and cannot directly reflect the relationship between users.

Summary of the invention

Embodiments of the present application provide a character relationship analysis method and related products, which can intuitively reflect the relationship between users and improve the efficiency of user relationship acquisition.

A first aspect of the embodiments of the present application provides a character relationship analysis method. The method includes:

Acquiring peer parameters between the target user and at least one peer user of the target user;

Establishing a target relationship graph between the target user and the at least one peer user according to the peer parameters of the at least one peer user;

The target user relationship between the target user and the at least one peer user is determined according to the target relationship graph.

Optionally, with reference to the first aspect of the embodiments of the present application, in a first possible implementation manner of the first aspect, the peer parameters include the number of peers and the peer distance, and the peers according to the at least one peer user Parameters to establish a target relationship graph between the target user and the at least one peer user, including:

Obtaining a first weight value of the number of peers, and a second weight value of the distance of the peers;

Weighting the number of peers of the at least one peer user and the peer distance of the at least one peer user by using the first weight value and the second weight value to obtain at least one target distance value;

Establish a reference relationship map with the target user's position point as the center of the circle and at least one target distance value as the radius, and the target user's position point as the target user's mapping point in the reference relationship map;

Obtaining the relationship graph correction factor of each peer user in the at least one peer user;

The target relationship graph is determined through the relationship graph correction factor of each peer user and the reference relationship graph.

Optionally, with reference to the first possible implementation manner of the first aspect of the embodiments of the present application, in a second possible implementation manner of the first aspect, the acquiring each of the at least one peer user Correction factors of the relationship graph of peer users, including:

Acquiring multiple images of a first peer user when traveling with the target user, the first peer user is any peer user of the at least one peer user;

Performing face recognition on the multiple images to obtain the face image of the first peer user and the face image of the target user in each of the multiple images;

Determine the angle between the face orientation of the first peer user and the target user in each image according to the face image of the first peer user in each image and the face image of the target user, to obtain Multiple target angles;

Determining the relationship graph correction factor of the first peer user according to the multiple target angles;

Adopting the acquisition method of the relationship graph correction factor of the first peer user to obtain the relationship graph correction factor of the second peer user to obtain the relationship graph correction factor of each peer user of the at least one peer user, the second The peer user is a user other than the first peer user among the at least one peer user.

A second aspect of an embodiment of the present application provides a character relationship analysis device. The device includes an acquisition unit, a creation unit, and a determination unit, where:

The acquiring unit is configured to acquire a peer parameter between the target user and at least one peer user of the target user;

The creating unit is configured to establish a target relationship graph between the target user and the at least one peer user according to the peer parameters of the at least one peer user;

The determining unit is configured to determine a target user relationship between the target user and the at least one peer user according to the target relationship graph.

Optionally, with reference to the second aspect of the embodiments of the present application, in a first possible implementation manner of the second aspect, the peer parameters include the number of peers and the distance of peers, based on the at least one peer user’s Peer parameters, to establish a target relationship graph between the target user and the at least one peer user, the creation unit is specifically used to:

Optionally, with reference to the first possible implementation manner of the second aspect of the embodiments of the present application, in the second possible implementation manner of the second aspect, each of the acquiring the at least one peer user With respect to the relationship graph correction factors of the same peer users, the creation unit is specifically used to:

A third aspect of the embodiments of the present application provides a terminal, including a processor, an input device, an output device, and a memory, where the processor, input device, output device, and memory are connected to each other, wherein the memory is used to store a computer program The computer program includes program instructions, and the processor is configured to call the program instructions to execute the step instructions in the first aspect of the embodiments of the present application.

A fourth aspect of the embodiments of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, wherein the computer program causes the computer to execute the first embodiment of the present application. Part or all of the steps described in one aspect.

A fifth aspect of an embodiment of the present application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium that stores the computer program, and the computer program is operable to cause the computer to execute as implemented in the present application Examples of some or all of the steps described in the first aspect. The computer program product may be a software installation package.

The implementation of the embodiments of the present application has at least the following beneficial effects:

Through the embodiment of the present application, the peer parameters between the target user and the at least one peer user of the target user are obtained, and between the target user and the at least one peer user are established according to the peer parameters of the at least one peer user Target relationship graph of the target relationship between the target user and the at least one peer user is determined according to the target relationship graph, therefore, relative to the existing solution, the relationship between users is based on existing data Analyze to get the user relationship. By analyzing the peer parameters of the target user and the target user's peer users, and establishing a relationship graph based on the peer parameters, the user relationship between the target user and the peer user can be intuitively reflected through the relationship graph Then, the user relationship between the target user and the peer users can be determined through the relationship graph, which can also improve the efficiency of obtaining user relationships to a certain extent.

BRIEF DESCRIPTION

In order to more clearly explain the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments of the present application. For those of ordinary skill in the art, without paying any creative labor, other drawings can be obtained based on these drawings.

1 is a schematic structural diagram of a character analysis system applying a character analysis method according to an embodiment of the present application;

2A is a schematic flowchart of a character relationship analysis method according to an embodiment of the present application;

2B is a schematic diagram of a reference relationship graph provided by an embodiment of the present application;

3 is a schematic flowchart of another character relationship analysis provided by an embodiment of the present application;

4 is a schematic flowchart of another character relationship analysis provided by an embodiment of the present application;

5 is a schematic structural diagram of a terminal according to an embodiment of the present application;

6 is a schematic structural diagram of a character relationship analysis device according to an embodiment of the present application.

detailed description

In order to better understand the character relationship analysis method provided by the embodiments of the present application, the following first briefly introduces the character analysis system applying the character analysis method. As shown in FIG. 1, the character analysis system includes a collection device 101 and an analysis device 102, wherein the collection device 101 collects peer parameters between the target user and at least one peer user of the target user, and the collection device 101 sends the peer parameters to the analysis device 102. The analyzing device 102 establishes a target relationship graph between the target user and at least one peer user according to the peer parameters of at least one peer user. After successfully establishing the target relationship graph, the analyzing device 102 determines the target user and the target user graph according to the target relationship graph The target user relationship between at least one peer user, therefore, compared with the existing solution, the relationship between users is analyzed through the existing data to obtain the user relationship, which can be passed by the target user and the target user's peer user's peer The parameters are analyzed, and a relationship graph is established according to the peer parameters, which can intuitively reflect the user relationship between the target user and the peer user through the relationship graph, and then the user relationship between the target user and the peer user can be determined through the relationship graph, thereby It can also improve the efficiency of acquiring user relationships to a certain extent.

Please refer to FIG. 2A, which is a schematic flowchart of a character relationship analysis method according to an embodiment of the present application. As shown in FIG. 2A, the character relationship analysis method includes steps 201-203, as follows:

201. Acquire a peer parameter between a target user and at least one peer user of the target user.

Optionally, the peer parameters may include the number of peers and the peer distance. The peer distance may be understood as the peer distance for each peer or the total distance when peers.

Optionally, the method for obtaining peer parameters may be to obtain the peer distance between the target user and each peer user in at least one peer user through a camera. Specifically, the target user may pass multiple cameras when traveling with the peer user. Acquire multiple images of peers separately, and determine the peer distance between the target user and peer users based on the multiple images of peers. A peer user can be understood as a user identified in the system and walking with the target user. Wherein, the multiple images are multiple images taken at preset time intervals, and the method for determining the peer distance between the target user and the peer user based on the multiple images includes steps A1-A4, as follows:

A1. According to each of the multiple images, determine the location point where the target user and the peer user are located, and obtain the movement speed of the target user and the peer user at each location point, including between two adjacent location points Multiple reference paths;

The reference path can be understood as a passable path between two adjacent positions. For example, if the two adjacent positions are points A and B, the reference path can be understood as between point A and point B. The path that users can pass.

A2. Determine the first reference distance according to the movement speed of the first position point and the preset time interval, and determine the second reference distance according to the movement speed of the second position point and the preset time interval, the first position point is The starting position, the second position is the ending position;

Among them, the preset time interval can be set by experience value or historical data.

Among them, the first position point and the second position point can be understood as, when the target user moves from the first position point to the second position point, the starting position point is the first position point, and the ending position point is the second position point, the first The position point may be any position point except the last position point among the plurality of position points, and the second position point is the next position point of the first position point.

A3. Obtain the distance between each reference path between the first location point and the second location point to obtain multiple third distances;

A4. The distance between the first reference distance and the second reference distance among the plurality of third distances is used as the peer distance between the target user and the peer user between the first location point and the second location point.

Optionally, if there are multiple third distances, the average of the multiple third distances is taken as the peer distance between the target user and the peer user between the first location point and the second location point.

In this example, the location distance and the movement speed of each location point are used to determine the peer user's peer distance between two adjacent location points, which can improve the accuracy of acquiring the peer distance to a certain extent.

202. Establish a target relationship graph between the target user and the at least one peer user according to the peer parameters of the at least one peer user.

Optionally, the peer parameters include peer distance and the number of peers. A possible method for establishing a target relationship graph based on peer parameters includes steps B1-B5, as follows:

B1. Obtain the first weight of the number of peers and the second weight of the distance of peers;

Among them, the first weight and the second weight can be obtained according to the neural network model. Specifically, the neural network model can be obtained by training as follows:

The neural network model can include an N-layer neural network. During training, the sample data can be input to the first layer of the N-layer neural network, and the first operation result can be obtained after the forward operation through the first layer, and then the first operation The result is input to the second layer to perform the forward operation to obtain the second result, and until the N-1 result is input to the Nth layer to perform the forward operation to obtain the Nth operation result, perform the reverse operation on the Nth operation result Training, so that the forward training and the reverse training are repeatedly performed until the neural network model training is completed. The sample data are the number of peers, the distance of peers, the preset first weight and the preset second weight.

B2. Using the first weight value and the second weight value to perform weight calculation on the number of peers of the at least one peer user and the peer distance of the at least one peer user to obtain at least one target distance value;

The weight calculation method may be: multiplying the first weight by the number of peers, multiplying the second weight by the distance of peers, and then summing to obtain a target distance value. The foregoing method is repeated until the target distance value of each peer user in at least one peer user is calculated to obtain at least one target distance value.

B3. Establish a reference relationship graph with the target user's location point as the center of the circle and at least one target distance value as the radius. The target user's location point is the target user's mapping point in the reference relationship graph;

Optionally, as shown in FIG. 2B, FIG. 2B is a schematic diagram of a reference relationship graph. Among them, r1, r2, r3 are different target distance values, and the center point is the mapping point of the target user in the reference relationship map.

B4. Obtain the relationship graph correction factor of each peer user in the at least one peer user;

Optionally, a possible method for obtaining the relationship graph correction factor of each peer user in at least one peer user includes steps B41-B45, as follows:

B41. Acquire multiple images of a first peer user when traveling with the target user, where the first peer user is any peer user of the at least one peer user;

Among them, multiple images can be collected by the camera. The camera may be a hyperspectral camera. When a hyperspectral camera is used, a possible collection method includes steps S1-S2, as follows:

S1: Collecting images of the first peer user and the target user when traveling together to obtain an image sequence including the target user and the peer user, and the image sequence includes multiple images;

Optionally, the hyperspectral camera can collect images in multiple bands, and when capturing images of target users and peer users, it can collect images in multiple different bands to obtain an image sequence including multiple bands . Multiple different bands can be represented as n bands, that is, from the first band to the n-th band, where n is a positive integer, where different bands can be understood as dividing a fixed band into multiple sub-bands, thereby obtaining multiple different bands The band, for example, the 300-600 micron band, which is divided into 10 bands, then 10 sub-bands can be obtained.

S2. Fusion the multiple images to obtain a target image.

Optionally, a method for fusing the multiple images includes steps S21-S22, as follows:

S21. Perform contour labeling on all pixels in each image in multiple images to obtain a labeling result. The labeling result includes human contour pixels and background pixels;

Wherein, the lower left corner of each image is taken as the origin of coordinates, the direction of the long side of the image is the x-axis, and the direction of the short side of the image is the y-axis to establish a rectangular coordinate system, thereby obtaining the coordinates of each pixel. Among them, the method of marking the body contour of the image can be marked according to the gray value of the pixel. When the gray value of the pixel is within the preset gray value interval, the pixel is marked as a human pixel, preset The gray value interval is set according to empirical values or according to historical data.

S22. Using the labeling result, fuse multiple images to obtain a target image.

According to the labeling result, the method for fusing the image may be: extract the labeling result of the pixel points of the same coordinates, if the labeling result of the coordinates of the same pixel point is that the number of human body pixels is greater than or equal to the preset number, then the point It is determined to be a pixel of the human body image. If the result of the marking of the coordinates of the same pixel point is that the number of human body pixels is lower than the preset number, the point is determined as the pixel of the background image; the target image is obtained according to the pixel of the image. The preset number is set according to empirical values or according to historical data.

By collecting human body images in multiple wave bands, and then merging human body images in multiple wave bands to obtain human body images, due to the different results of imaging different colors in different wave bands, for example, when users wear Wearing different colors of clothes, etc., the imaging results of different wave bands can better reflect the characteristics of the human body, so the multi-band human body image fusion method is used to obtain the human body image, which can improve the accuracy of obtaining the human body image to a certain extent.

B42. Perform face recognition on the multiple images to obtain the face image of the first peer user and the face image of the target user in each of the multiple images;

Optionally, when the face image is occluded, a possible multiple images are analyzed to determine the face image of the first peer user and the face image of the target user include steps B420-B428, details as follows:

B420. Repair the target face image according to the principle of face symmetry to obtain the first face image and the target repair coefficient. The target repair coefficient is used to express the completeness of the face image for repair;

Alternatively, the target face image can be mirror-transformed according to the principle of face symmetry. After the mirror-transform process, the target face image after processing can be used to perform face repair based on the model of generating an adversarial network to obtain The first face image and the target repair coefficient.

Wherein, the target face image is a face image extracted from the collected image and including only part of the face. The target repair coefficient may be the proportion of pixels of the face part of the repaired face to the total number of pixels of the entire face. The model for generating the adversarial network may include the following components: discriminator, semantic regularization network, etc., which are not limited here. .

B421. Perform feature extraction on the first face image to obtain a first face feature set, and perform feature extraction on the target face image to obtain a second face feature set;

Optionally, the method of feature extraction on the first face image may include at least one of the following: LBP (Local Binary Patterns) feature extraction algorithm, HOG (Histogram of Oriented Gradient, directional gradient histogram) feature Extraction algorithms, LoG (Laplacian of Gaussian, second-order Laplacian-Gaussian) feature extraction algorithms, etc., are not limited here.

B422. Perform a search in the database according to the first face feature set to obtain face images of multiple objects that successfully match the first face feature set;

Among them, the face feature set is pre-stored in the database.

B423. Match the second face feature set with the feature sets of the face images of the multiple objects to obtain multiple first matching values;

B424. Obtain the human body characteristic data of each of the multiple objects to obtain multiple human body characteristic data;

B425. Match the human body feature data corresponding to the target face with each of the multiple human body feature data to obtain multiple second matching values;

B426. Determine a first weight corresponding to the target repair coefficient according to a preset mapping relationship between the repair coefficient and the weight, and determine a second weight according to the first weight;

B427. Perform a weighted operation according to the first weight value, the second weight value, the multiple first matching values, and the multiple second matching values to obtain multiple target matching values;

B428. Select the maximum value from the multiple target matching values, and use the object corresponding to the maximum value as the complete face image corresponding to the target face image.

The mapping relationship between the preset repair coefficients and the weights may be that each preset repair coefficient corresponds to a weight, and the sum of the weights of each preset repair coefficient is 1, the preset The weight of the repair coefficient can be set by the user or the system default. Specifically, according to the preset mapping relationship between the repair coefficient and the weight, the first weight corresponding to the target repair coefficient is determined, and according to the first weight The value determines the second weight value. The second weight value may be the weight value corresponding to the second matching value. The sum between the first weight value and the second weight value is 1, and the first weight value is matched with multiple first weights respectively. Value weighting, and weighting the second weights with multiple second matching values respectively to obtain multiple target matching values corresponding to multiple objects, and selecting the object corresponding to the largest matching value among the multiple matching values as the target person The complete face image corresponding to the face image.

In this example, by repairing the incomplete face image, matching the repaired face image to obtain the face images of multiple objects, and by comparing the human features, the target face image is determined Corresponding to the complete face image, through the repair of the face, the matched image after the repair is filtered to obtain the final complete face image, which can improve the accuracy of the face image acquisition to a certain extent.

B43. Determine the angle between the face orientation of the first peer user and the target user in each image based on the face image of the first peer user and the face image of the target user in each image To get multiple target angles;

The angle between the face orientation of the first peer user and the face orientation of the target user can be understood as abstracting the face as a plane, the direction of the vertical line of the plane is the face orientation, and the direction of the vertical line is the face Direction, therefore, the angle between the face orientation of the first peer user and the face orientation of the target user is the angle between the vertical lines. The angle between vertical lines can be understood as the angle between straight lines of different planes or the angle between coplanar straight lines.

B44. Determine the relationship graph correction factor of the first peer user according to the multiple target angles;

Among them, the average value of the plurality of target angles is calculated to obtain the average value of the target angle, and then the relationship of the first peer user corresponding to the average value of the target angle is determined according to the mapping relationship between the average angle and the correction factor of the relationship map The map correction factor, the mapping between the mean angle and the relationship map correction factor is stored in the character relationship analysis system in advance. The relationship graph correction factor can be understood as a parameter that corrects the target distance value.

B45. Use the acquisition method of the relationship graph correction factor of the first peer user to obtain the relationship graph correction factor of the second peer user, and obtain the relationship graph correction factor of each peer user of the at least one peer user. The second peer user is a user other than the first peer user among the at least one peer user.

Wherein, the way to obtain the relationship graph correction factor of the first peer user can be understood as the method described in steps B41-B44.

B5. Determine the target relationship graph by using the relationship graph correction factor of each peer user and the reference relationship graph.

Optionally, the relationship graph correction factor may be multiplied by the corresponding target distance value in the reference relationship graph to obtain the target relationship graph.

203. Determine a target user relationship between the target user and the at least one peer user according to the target relationship graph.

Optionally, the target relationship graph can reflect the intimacy between peer users and target users. The smaller the target distance value, the higher the intimacy, and the larger the target distance value, the lower the intimacy, with the same target distance value , The same intimacy.

In a possible example, in order to improve the security of the target user, the security of the users traveling with the target user can be analyzed through the target relationship graph. This method can be treasure steps C1-C3, as follows:

C1, at least one reference peer user is determined from the at least one peer user, and the reference peer user is a peer user whose intimacy is lower than a preset intimacy;

Among them, the method of obtaining the intimacy of the reference peer users can be obtained from the generated target relationship graph. The preset intimacy can be set by experience value setting or historical data.

C2. Obtain the peer distance between the target user and each of the at least one reference peer user;

Optionally, the peer distance may be the peer distance when the target user and the reference peer user are currently traveling. The method for obtaining the peer distance between the target user and the reference user can refer to the method for determining the peer distance in steps A1 to A4, which will not be described here.

C3. If there is at least one target reference peer user, the alarm information is pushed to the target user. The target reference peer user is a user whose peer distance among the at least one reference peer user is greater than a preset peer distance.

Optionally, because the distance between the target user and the reference peer user is longer and the intimacy is lower, it can be determined that the reference peer user may pose a security risk to the target user, and if it is a potentially dangerous user, an alarm can be pushed to the target user Information, dangerous users can be understood as users who pose a threat to the personal property of others, for example, they will deceive the target user's money through some bad means, directly rob the target user, etc.

Optionally, a possible method for pushing the alarm information to the target user includes steps C31-C34, as follows:

C31. Obtain the action information of the target user;

The motion information may include foot motion and head motion. Head movements can include normal state, shaking state, and head tilt state. The normal state can be understood as the state of normal walking, the shaking state can be understood as left and right shaking or back and forth shaking, etc. The shaking state includes shaking frequency, and the eccentric head state can be understood The head is tilted to the left or right. Foot movements can include pace, pace direction, etc. The pace direction can be understood as the walking direction, and the pace can be understood as the distance of each step and the frequency of walking.

C32. Determine the state information of the target user according to the action information;

Optionally, the user state may include: a normal state and a bad state. The bad state may include a mild bad state, a moderate bad state, and a severe bad state. The mild bad state may be understood as that the target user's consciousness is in a slightly blurred state. For example, the mild drunk state, the mild drunk state can be understood as the target user can walk normally but the consciousness is slightly blurred, and the moderate bad state can be understood as the target user's consciousness is in a moderately blurred state, for example, moderate drunk state, moderate The drunk state can be understood as that the target user can no longer walk normally, but can still walk. The severely bad state can be understood as the target user's consciousness is in a severely blurred state, for example, the severe drunk state, the severe drunk state can be understood as that the target user has If you cannot walk normally, you will be accompanied by vomiting.

Optionally, the method for determining the state information of the target user according to the action information can be determined by the mapping relationship between the action information and the state information. A possible mapping relationship between the action information and the state information is shown in Table 1. , Taking head movement as an example, as follows:

Table 1

动作信息Action information	用户状态user status
头部正常状态Normal state of the head	正常状态normal status
晃动状态(晃动频率低)Shaking state (low shaking frequency)	中度不良状态Moderately bad condition
晃动状态(晃动频率高)Shaking state (high shaking frequency)	轻度不良状态Mildly bad condition
偏头状态Biased state	中度不良状态或重度不良状态Moderately bad state or severely bad state

Here, only the head motion is taken as an example for description, and of course, there may be a state information mapping relationship combining head motion and foot motion.

C33. Determine the target security level of the target user according to the status information;

Optionally, the target security level corresponds to the user status one by one, and the security level in the normal status is the highest, followed by a mildly bad status, a moderately poor status, and a severely poor status, and the security level is sequentially lowered.

C34. If the target security level of the target user is lower than a preset security level, then push alarm information corresponding to the target security level to the target user at a preset time interval.

Among them, the preset security level is the security level corresponding to the mild bad state, and the preset time interval can be set by experience value or historical data. The alarm information corresponding to the target security level can be understood as that different security levels have different alarm information, and the specific alarm information is set by empirical values.

In this example, the target user's own status is determined and the target user's security level is determined based on the status. When the security level is low, the target user is warned. For example, the target user goes home alone after drinking. When the security level of the target user is low, the security level of the target user needs to be discriminated, and when the security level of the target user is determined to be low, an alarm message is sent to the user, which can alert the target user, thereby improving the target user to a certain extent Security.

Please refer to FIG. 3, which is a schematic flowchart of another character relationship analysis provided by an embodiment of the present application. As shown in FIG. 3, the character relationship analysis method may include steps 301-307, as follows:

301: Obtain a peer parameter between a target user and at least one peer user of the target user;

Wherein, the peer parameters include the number of peers and the distance of peers, according to the peer parameters of the at least one peer user.

302. Obtain a first weight value of the number of peers and a second weight value of the distance of the peers;

303. Perform weight calculation on the number of peers of the at least one peer user and the peer distance of the at least one peer user by using the first weight and the second weight to obtain at least one target distance value;

304. Establish a reference relationship graph with the target user's location point as the center of the circle and at least one target distance value as the radius. The target user's location point is the target user's mapping point in the reference relationship graph;

305. Obtain a relationship graph correction factor for each peer user in the at least one peer user;

Wherein, the method for obtaining the correction factor of the relationship graph may be: acquiring multiple images when the first peer user is traveling with the target user, and the first peer user is any one of the at least one peer user; Performing face recognition on the multiple images to obtain the face image of the first peer user and the face image of the target user in each of the multiple images; according to the first peer in each image The face image of the user and the face image of the target user, determining the angle between the face orientation of the first peer user and the target user in each image to obtain multiple target angles; according to the multiple Target angles to determine the relationship graph correction factor of the first peer user; using the acquisition method of the relationship graph correction factor of the first peer user to obtain the relationship graph correction factor of the second peer user to obtain the at least one A relationship graph correction factor of each peer user among peer users, the second peer user is a user other than the first peer user among the at least one peer user.

306. Determine the target relationship graph by using the relationship graph correction factor of each peer user and the reference relationship graph;

307. Determine a target user relationship between the target user and the at least one peer user according to the target relationship graph.

In this example, the target distance value is determined according to the number of peers and the distance of the peers, and the reference relationship graph is generated according to the target distance value. After the reference relationship graph is corrected, the target relationship graph is obtained, which can improve the acquisition of the target relationship graph to a certain extent. Accuracy.

Please refer to FIG. 4, which is a schematic flowchart of another character relationship analysis provided by an embodiment of the present application. As shown in FIG. 4, the character relationship analysis method may include steps 401-406, as follows:

401. Acquire peer parameters between a target user and at least one peer user of the target user;

402. Establish a target relationship graph between the target user and the at least one peer user according to the peer parameters of the at least one peer user;

403. Determine a target user relationship between the target user and the at least one peer user according to the target relationship graph;

404. Determine at least one reference peer user from the at least one peer user, where the reference peer user is a peer user whose intimacy is lower than a preset intimacy;

405. Obtain the peer distance between the target user and each of the at least one reference peer user;

406. If the travel distance of one of the at least one reference peer user is higher than a preset distance, acquire the action information of the target user;

407. Determine the state information of the target user according to the action information;

408. Determine the target security level of the target user according to the status information;

409. If the target security level of the target user is lower than a preset security level, push alarm information corresponding to the target security level to the target user at a preset time interval.

Consistent with the above embodiment, please refer to FIG. 5, which is a schematic structural diagram of a terminal provided by an embodiment of the present application. As shown in the figure, it includes a processor, an input device, an output device, and a memory. The input device, the output device, and the memory are connected to each other, wherein the memory is used to store a computer program, the computer program includes program instructions, the processor is configured to call the program instructions, and the above program includes to execute the following Step instructions

In this example, the peer parameters between the target user and at least one peer user of the target user are obtained, and the target between the target user and the at least one peer user is established according to the peer parameters of the at least one peer user Relationship graph, the target user relationship between the target user and the at least one peer user is determined according to the target relationship graph, therefore, relative to the existing solution, the relationship between users is analyzed through existing data To get the user relationship, by analyzing the peer parameters of the target user and the target user’s peer users, and establishing a relationship graph based on the peer parameters, the user relationship between the target user and peer users can be intuitively reflected through the relationship graph, and then The user relationship between the target user and the peer users can be determined through the relationship graph, which can also improve the efficiency of obtaining the user relationship to a certain extent.

The above mainly introduces the solutions of the embodiments of the present application from the perspective of the execution process on the method side. It can be understood that, in order to realize the above-mentioned functions, the terminal includes a hardware structure and/or a software module corresponding to each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of the examples described in the embodiments provided herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is performed by hardware or computer software driven hardware depends on the specific application of the technical solution and design constraints. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of this application.

The embodiments of the present application may divide the functional unit of the terminal according to the above method example, for example, each functional unit may be divided corresponding to each function, or two or more functions may be integrated into one processing unit. The above integrated unit can be implemented in the form of hardware or software function unit. It should be noted that the division of the units in the embodiments of the present application is schematic, and is only a division of logical functions, and there may be other division manners in actual implementation.

Consistent with the above, please refer to FIG. 6, which is a schematic structural diagram of an apparatus for analyzing a person relationship according to an embodiment of the present application. The apparatus includes an obtaining unit 601, a creating unit 602, and a determining unit 603, where,

The obtaining unit 601 is configured to obtain a peer parameter between the target user and at least one peer user of the target user;

The creating unit 602 is configured to establish a target relationship graph between the target user and the at least one peer user according to the peer parameters of the at least one peer user;

The determining unit 603 is configured to determine a target user relationship between the target user and the at least one peer user according to the target relationship graph.

Optionally, the peer parameters include the number of peers and the peer distance. In terms of establishing a target relationship graph between the target user and the at least one peer user based on the peer parameters of the at least one peer user, all The creation unit 602 is specifically used for:

Optionally, in terms of acquiring the relationship graph correction factor of each peer user of the at least one peer user, the creating unit 602 is specifically configured to:

Determine the angle between the face orientation of the first peer user and the target user in each image according to the face image of the first peer user and the face image of the target user in each image, to obtain Multiple target angles;

The above method is repeated until the relationship graph correction factor of each peer user in the at least one peer user is obtained.

Optionally, the character relationship analysis device is also specifically used for:

Determining at least one reference peer user from the at least one peer user, the reference peer user being a peer user whose intimacy is lower than a preset intimacy;

Obtaining the peer distance between the target user and each of the at least one reference peer user;

If there is at least one target reference peer user, alarm information is pushed to the target user, and the target reference peer user is a user whose peer distance among the at least one reference peer user is greater than a preset peer distance.

Optionally, in terms of pushing the alarm information to the target user, the person relationship analysis device is further specifically used to:

Obtain the action information of the target user;

Determine the status information of the target user according to the action information;

Determine the target security level of the target user according to the status information;

If the target security level of the target user is lower than the preset security level, the alarm information corresponding to the target security level is pushed to the target user at a preset time interval.

An embodiment of the present application further provides a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, and the computer program causes the computer to execute any of the character relationship dividing methods described in the above method embodiments Part or all steps.

An embodiment of the present application further provides a computer program product, the computer program product includes a non-transitory computer-readable storage medium storing a computer program, the computer program causes the computer to execute any of the characters described in the above method embodiments Part or all of the steps of the relationship analysis method.

Claims

A character relationship analysis method, characterized in that the method includes:

Acquiring peer parameters between the target user and at least one peer user of the target user;

Establishing a target relationship graph between the target user and the at least one peer user according to the peer parameters of the at least one peer user;

The target user relationship between the target user and the at least one peer user is determined according to the target relationship graph.
The method according to claim 1, wherein the peer parameters include a number of peers and a peer distance, and based on the peer parameters of the at least one peer user, establishing a relationship between the target user and the at least one peer user Atlas of target relationships, including:

Obtaining a first weight value of the number of peers, and a second weight value of the distance of the peers;

Weighting the number of peers of the at least one peer user and the peer distance of the at least one peer user by using the first weight value and the second weight value to obtain at least one target distance value;

Establish a reference relationship map with the target user's position point as the center of the circle and at least one target distance value as the radius, and the target user's position point as the target user's mapping point in the reference relationship map;

Obtaining the relationship graph correction factor of each peer user in the at least one peer user;

The target relationship graph is determined through the relationship graph correction factor of each peer user and the reference relationship graph.
The method according to claim 2, wherein the acquiring the relationship graph correction factor of each peer user of the at least one peer user comprises:

Acquiring multiple images of a first peer user when traveling with the target user, the first peer user is any peer user of the at least one peer user;

Performing face recognition on the multiple images to obtain the face image of the first peer user and the face image of the target user in each of the multiple images;

Determine the angle between the face orientation of the first peer user and the target user in each image according to the face image of the first peer user and the face image of the target user in each image, to obtain Multiple target angles;

Determining the relationship graph correction factor of the first peer user according to the multiple target angles;

Adopting the acquisition method of the relationship graph correction factor of the first peer user to obtain the relationship graph correction factor of the second peer user to obtain the relationship graph correction factor of each peer user of the at least one peer user, the second The peer user is a user other than the first peer user among the at least one peer user.
The method according to claim 2, wherein the target user relationship includes intimacy, and the method further includes:

Determining at least one reference peer user from the at least one peer user, the reference peer user being a peer user whose intimacy is lower than a preset intimacy;

Obtaining the peer distance between the target user and each of the at least one reference peer user;

If there is at least one target reference peer user, alarm information is pushed to the target user, and the target reference peer user is a user whose peer distance among the at least one reference peer user is greater than a preset peer distance.
The method according to claim 4, wherein the pushing of the alarm information to the target user comprises:

Obtain the action information of the target user;

Determine the status information of the target user according to the action information;

Determine the target security level of the target user according to the status information;

If the target security level of the target user is lower than the preset security level, the alarm information corresponding to the target security level is pushed to the target user at a preset time interval.
A character relationship analysis device, characterized in that the device includes an acquisition unit, a creation unit and a determination unit, wherein,

The obtaining unit is configured to obtain a peer parameter between the target user and at least one peer user of the target user;

The creating unit is configured to establish a target relationship graph between the target user and the at least one peer user according to the peer parameters of the at least one peer user;

The determining unit is configured to determine a target user relationship between the target user and the at least one peer user according to the target relationship graph.
The apparatus according to claim 6, wherein the peer parameters include a number of peers and a peer distance, and the target user and the at least one peer user are established based on the peer parameters of the at least one peer user In terms of the target relationship graph, the creation unit is specifically used to:

Obtaining a first weight value of the number of peers, and a second weight value of the distance of the peers;

Weighting the number of peers of the at least one peer user and the peer distance of the at least one peer user by using the first weight value and the second weight value to obtain at least one target distance value;

Establish a reference relationship map with the target user's position point as the center of the circle and at least one target distance value as the radius, and the target user's position point as the target user's mapping point in the reference relationship map;

Obtaining the relationship graph correction factor of each peer user in the at least one peer user;

The target relationship graph is determined through the relationship graph correction factor of each peer user and the reference relationship graph.
The apparatus according to claim 7, characterized in that, in terms of acquiring the relationship graph correction factor of each peer user of the at least one peer user, the creating unit is specifically configured to:

Acquiring multiple images of a first peer user when traveling with the target user, the first peer user is any peer user of the at least one peer user;

Performing face recognition on the multiple images to obtain the face image of the first peer user and the face image of the target user in each of the multiple images;

Determine the angle between the face orientation of the first peer user and the target user in each image according to the face image of the first peer user in each image and the face image of the target user, to obtain Multiple target angles;

Determining the relationship graph correction factor of the first peer user according to the multiple target angles;

Adopting the acquisition method of the relationship graph correction factor of the first peer user to obtain the relationship graph correction factor of the second peer user to obtain the relationship graph correction factor of each peer user of the at least one peer user, the second The peer user is a user other than the first peer user among the at least one peer user.
A terminal is characterized by comprising a processor, an input device, an output device, and a memory, where the processor, input device, output device, and memory are connected to each other, wherein the memory is used to store a computer program, and the computer program Including program instructions, the processor is configured to call the program instructions to perform the method according to any one of claims 1-5.
A computer-readable storage medium, characterized in that the computer storage medium stores a computer program, and the computer program includes program instructions, which when executed by a processor causes the processor to execute as claimed in claim 1 -5 The method of any one.