Scene graph recommendation method and device based on model graph, equipment and medium

Abstract

The invention provides a scene graph recommendation method and device based on a model graph, equipment and a medium, and the method comprises the steps: setting a scene picture library, setting label information for each scene graph in the scene picture library, and enabling the label information to comprise a gender label, an age label, a figure integrity label and a scene label; performing label setting on the uploaded model chart, wherein the label setting comprises a model gender label, a model age label, a model figure integrity label and a model scene label; performing label matching on a model gender label, a model age label, a model figure integrity label and a model scene label with each scene picture in the scene picture library, and calculating a matching degree; according to the sequence of the matching degrees from high to low, arranging and displaying the scene graphs; and the scene graph recommendation accuracy is greatly improved.

Description

Technical Field

[0001] The present invention relates to a method, device, equipment and medium for recommending a scene graph based on a model graph. Background Art

[0002] In the prior art, an AI model replacement function is provided. By uploading a model graph by a user and then providing a large number of scene reference graphs, after the user selects a scene reference graph, the model in the model graph uploaded by the user can be transferred to the selected scene reference graph. However, the existing scene reference graphs are only simply classified, and then the user selects one of the classifications and then views them one by one to select the required reference graph. Therefore, the user needs to browse a lot of scene graphs to find a scene graph that truly matches the uploaded model graph and meets the user's needs. This method is inefficient. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method, device, equipment and medium for recommending a scene graph based on a model graph. By matching the uploaded picture with the scene graph and combining gender and age, very accurate scene graph recommendation can be achieved, greatly saving the user's time and improving work efficiency.

[0004] In a first aspect, the present invention provides a method for recommending a scene graph based on a model graph, including the following steps:

[0005] Step 1: Set up a scene picture library, and set tag information for each scene graph in the scene picture library. The tag information includes a gender tag, an age tag, a character integrity tag, and a scene tag;

[0006] Step 2: Set tags for the uploaded model graph, including a model gender tag, a model age tag, a model character integrity tag, and a model scene tag;

[0007] Step 3: Respectively match the model gender tag, the model age tag, the model character integrity tag, and the model scene tag with each scene graph in the scene picture library, and calculate the matching degree; arrange and display the scene graphs in descending order of the matching degree.

[0008] In a second aspect, the present invention provides a device for recommending a scene graph based on a model graph, including:

[0009] A picture library setting module, which sets up a scene picture library and sets tag information for each scene graph in the scene picture library. The tag information includes a gender tag, an age tag, a character integrity tag, and a scene tag;

[0010] Set up a model image tagging module to tag the uploaded model images, including model gender tags, model age tags, model integrity tags, and model scene tags;

[0011] A comparison and recommendation module that matches the model gender tags, model age tags, model integrity tags, and model scene tags with each scene image in the scene image library respectively, calculates the matching degree, and arranges and displays the scene images in descending order of the matching degree.

[0012] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the method described in the first aspect is implemented.

[0013] In a fourth aspect, the present invention provides a computer-readable storage medium with a computer program stored thereon. When the program is executed by a processor, the method described in the first aspect is implemented.

[0014] One or more technical solutions provided by the present invention have at least the following technical effects or advantages:

[0015] The present invention makes a very accurate recommendation of scene images by matching the uploaded images with scene images and combining gender and age, greatly saving the user's time and improving work efficiency; the user can use the existing algorithm to combine the scene images with the uploaded images to generate the required images.

[0016] The above description is only an overview of the technical solution of the present invention. In order to be able to understand the technical means of the present invention more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention are given below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below with reference to the accompanying drawings in conjunction with embodiments.

[0018] Figure 1 It is the flowchart in the method of the first embodiment of the present invention;

[0019] Figure 2 It is the structural schematic diagram of the device in the second embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The overall idea of the technical solution in the embodiments of the present application is as follows:

[0021] Background preparation:

[0022] (1) There are four built-in tags for model images:

[0023] Gender labels: male, female;

[0024] Character integrity labels: upper body picture with 5 points, upper body picture with 7 points, upper body picture with 9 points, lower body picture, close-up picture of legs and feet, first full body picture occupying the screen, second full body picture occupying the screen, third full body picture occupying the screen, fourth full body picture occupying the screen, fifth full body picture occupying the screen, and close-up picture of the head;

[0025] Age labels: infant, child, youth, middle-aged, elderly;

[0026] Scene labels: playground, living room, bedroom, kitchen, etc., nearly a thousand kinds;

[0027] (2) Scene picture library, the scene picture library includes multiple scene pictures, and the gender label, age label, and character integrity label of the scene picture are identified through existing recognition algorithms; then, through task integrity classification, the character integrity label is determined;

[0028] Recommended pictures:

[0029] (1) For the model picture uploaded by the user, identify the belonging of the four corresponding labels of the model picture; among them, the age label, scene label, and gender label can all be empty;

[0030] (2) Perform label matching for the four types of labels in the scene picture library respectively:

[0031] One hit of the gender label adds x points, and it can hit at most once;

[0032] One hit of the character integrity label adds y points, and it can hit at most once;

[0033] One hit of the age label adds y points, and it can hit at most once;

[0034] One hit of the scene label adds z points, and it can hit at most five times. If it exceeds five times, it is also calculated as five times;

[0035] The above scores: x > y > 5z. Among them, the age, character integrity, and gender of the model are relatively crucial. The scene only has a little influence on the recommended ranking when the scores of age, character integrity, and gender are the same. Because if the user uploads a real-shot picture of the model, the background may just be a white wall, so the scene matching actually has no meaning; among these three, age and character integrity are more important than gender;

[0036] (3) After accumulating various labels, finally calculate the score of each model picture in the library in this round of matching, and sort them in reverse order according to the score. The most suitable scene for the user is ranked at the front, and the order between the scene pictures with the same score is randomly arranged, so that the user can quickly make a choice.

[0037] Assume that hitting the gender label adds 10 points, hitting the character integrity label adds 20 points, hitting the age label adds 20 points, and hitting the scene label adds 1 point with a ceiling of 5 points; then the matching degree can be calculated.

[0038] Example 1

[0039] As Figure 1 shown, this embodiment provides a method for recommending a scene graph based on a model graph, including the following steps:

[0040] Step 1: Set up a scene picture library, and set label information for each scene picture in the scene picture library. The label information includes a gender label, an age label, a character integrity label, and a scene label;

[0041] Step 2: Set labels for the uploaded model graph, including a model gender label, a model age label, a model character integrity label, and a model scene label;

[0042] Step 3: Match the model gender label, the model age label, the model character integrity label, and the model scene label with each scene picture in the scene picture library respectively to calculate the matching degree; arrange and display the scene pictures in descending order of the matching degree; and the user can then make a selection.

[0043] In this embodiment, preferably, the character integrity label includes: a 5-point upper body picture, a 7-point upper body picture, a 9-point upper body picture, a lower body picture, a close-up of legs, a first full body picture occupying the screen, a second full body picture occupying the screen, a third full body picture occupying the screen, a fourth full body picture occupying the screen, a fifth full body picture occupying the screen, and a close-up of the head; the judgment method of the character integrity label is:

[0044] Obtain the model graph uploaded by the user, and obtain the skeleton graph of the model graph; the skeleton graph includes key point information;

[0045] 5-point upper body picture:

[0046] If at least one of the right hip key point and the left hip key point exists, and the right hip key point or the left hip key point is less than 20 pixels away from the bottom of the skeleton graph, and both the right knee key point and the left knee key point do not exist; then it is a 5-point upper body picture;

[0047] If both the left shoulder key point and the right shoulder key point exist, and the distance between the left shoulder key point and the right shoulder key point is less than 0.7 of the width of the skeleton graph, and the right eye key point, the left eye key point, the right ear key point, the left ear key point, and the nose key point all exist, and the distance between the right ear key point and the left ear key point is less than 1 / 3 of the width of the skeleton graph; then it is a 5-point upper body picture;

[0048] 7-point upper body picture:

[0049] If both the right hip key point and the left hip key point exist, at least one of the right knee key point and the left knee key point exists, and the right knee key point or the left knee key point is less than 20 pixels away from the bottom of the skeleton diagram, and both the right ankle key point and the left ankle key point do not exist, it is a 7-point upper body diagram;

[0050] 9-point upper body diagram:

[0051] If both the right hip key point and the left hip key point exist, both the right knee key point and the left knee key point exist, at least one of the right ankle key point and the left ankle key point exists, and the right ankle key point or the left ankle key point is less than 20 pixels away from the bottom of the skeleton diagram, it is a 9-point upper body diagram;

[0052] Lower body diagram:

[0053] If the neck key point does not exist, both the right hip key point and the left hip key point exist, both the right knee key point and the left knee key point exist, and both the right ankle key point and the left ankle key point exist, it is a lower body diagram;

[0054] Leg and foot close-up diagram:

[0055] If the neck key point does not exist, both the right hip key point and the left hip key point do not exist, both the right knee key point and the left knee key point exist, and both the right ankle key point and the left ankle key point exist, it is a leg and foot close-up diagram;

[0056] First full body occupying the picture diagram:

[0057] If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 80% of the height of the skeleton diagram, it is the first full body occupying the picture diagram;

[0058] Second full body occupying the picture diagram:

[0059] If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 70% and less than or equal to 80% of the height of the skeleton diagram, it is the second full body occupying the picture diagram;

[0060] Third full body occupying the picture diagram:

[0061] If the right-eye key point, left-eye key point, right-ear key point, left-ear key point, nose key point, left-ankle key point, and right-ankle key point all exist, and the ratio of the distance between the left-ankle key point and the left-knee key point to the distance between the left-knee key point and the left-hip key point is within the set threshold, and the distance from the nose key point to the right-ankle key point is greater than 60% of the height of the skeleton diagram and less than or equal to 70% of the height of the skeleton diagram, then it is the third full-body picture occupying the screen;

[0062] Fourth full-body picture occupying the screen:

[0063] If the right-eye key point, left-eye key point, right-ear key point, left-ear key point, nose key point, left-ankle key point, and right-ankle key point all exist, and the ratio of the distance between the left-ankle key point and the left-knee key point to the distance between the left-knee key point and the left-hip key point is within the set threshold, and the distance from the nose key point to the right-ankle key point is greater than 50% of the height of the skeleton diagram and less than or equal to 60% of the height of the skeleton diagram, then it is the fourth full-body picture occupying the screen;

[0064] Fifth full-body picture occupying the screen:

[0065] If the right-eye key point, left-eye key point, right-ear key point, left-ear key point, nose key point, left-ankle key point, and right-ankle key point all exist, and the ratio of the distance between the left-ankle key point and the left-knee key point to the distance between the left-knee key point and the left-hip key point is within the set threshold, and the distance from the nose key point to the right-ankle key point is less than or equal to 50% of the height of the skeleton diagram, then it is the fifth full-body picture occupying the screen;

[0066] Close-up picture of the head:

[0067] If the right-eye key point, left-eye key point, right-ear key point, left-ear key point, and nose key point all exist and it does not belong to the 5-point upper-body picture, then they are all close-up pictures of the head;

[0068] Make judgments in sequence. If it meets the requirements, end the judgment and label it.

[0069] In this embodiment, preferably, step 3 is specifically as follows: Match the model gender label, model age label, model task label, and model scene label with each scene picture in the scene picture library respectively to calculate the matching degree; Arrange and display the scene pictures in descending order of the matching degree;

[0070] If the gender label of the scene picture is the same as the model gender label, add x points; otherwise, add 0 points;

[0071] If the character integrity label of the scene picture is the same as the model character integrity label, add y points; otherwise, add 0 points;

[0072] If the age label of the scene graph is the same as the age label of the model, add z points; otherwise, add 0 points;

[0073] Compare the scene label of the scene graph with the scene label, and calculate and add m*n points according to the number m of matches, where 5≥m≥0; x>y>z>n; Each scene graph includes at least one scene label, and each model graph includes at least one scene label;

[0074] The matching degree is the sum of the above scores.

[0075] Based on the same inventive concept, the present application also provides a device corresponding to the method in Embodiment 1. For details, see Embodiment 2.

[0076] Embodiment 2

[0077] As Figure 2 shown, in this embodiment, a scene graph recommendation device based on a model graph is provided, including:

[0078] A gallery module is set up to set up a scene picture library, and label information is set for each scene graph in the scene picture library. The label information includes a gender label, an age label, a character integrity label, and a scene label;

[0079] A model graph label module is set up to set labels for the uploaded model graph, including a model gender label, a model age label, a model character integrity label, and a model scene label;

[0080] A comparison and recommendation module is used to perform label matching on the model gender label, the model age label, the model character integrity label, and the model scene label with each scene graph in the scene picture library respectively, and calculate the matching degree; According to the order from high to low of the matching degree, the scene graphs are arranged and displayed.

[0081] In this embodiment, preferably, the character integrity label includes: 5-point upper body diagram, 7-point upper body diagram, 9-point upper body diagram, lower body diagram, leg close-up diagram, first full body occupying the picture diagram, second full body occupying the picture diagram, third full body occupying the picture diagram, fourth full body occupying the picture diagram, fifth full body occupying the picture diagram, and head close-up diagram; The judgment method of the character integrity label is:

[0082] Obtain the model graph uploaded by the user, and obtain the skeleton graph of the model graph; The skeleton graph includes key point information;

[0083] 5-point upper body diagram:

[0084] If at least one of the right hip key point and the left hip key point exists, and the right hip key point or the left hip key point is less than 20 pixels away from the bottom of the skeleton graph, and both the right knee key point and the left knee key point do not exist; then it is a 5-point upper body diagram;

[0085] If both the left shoulder key point and the right shoulder key point exist, and the distance between the left shoulder key point and the right shoulder key point is less than 0.7 of the width of the skeleton diagram, and the right eye key point, the left eye key point, the right ear key point, the left ear key point, and the nose key point all exist, and the distance between the right ear key point and the left ear key point is less than 1 / 3 of the width of the skeleton diagram; then it is a 5-point upper body diagram;

[0086] 7-point upper body diagram:

[0087] If both the right hip key point and the left hip key point exist, at least one of the right knee key point and the left knee key point exists, and the right knee key point or the left knee key point is less than 20 pixels away from the bottom of the skeleton diagram, and neither the right ankle key point nor the left ankle key point exists, then it is a 7-point upper body diagram;

[0088] 9-point upper body diagram:

[0089] If both the right hip key point and the left hip key point exist, both the right knee key point and the left knee key point exist, at least one of the right ankle key point and the left ankle key point exists, and the right ankle key point or the left ankle key point is less than 20 pixels away from the bottom of the skeleton diagram, then it is a 9-point upper body diagram;

[0090] Lower body diagram:

[0091] If the neck key point does not exist, both the right hip key point and the left hip key point exist, both the right knee key point and the left knee key point exist, and both the right ankle key point and the left ankle key point exist, then it is a lower body diagram;

[0092] Close-up of legs and feet diagram:

[0093] If the neck key point does not exist, neither the right hip key point nor the left hip key point exists, both the right knee key point and the left knee key point exist, and both the right ankle key point and the left ankle key point exist, then it is a close-up of legs and feet diagram;

[0094] First full body occupying the picture diagram:

[0095] If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point, and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance from the left knee key point to the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 80% of the height of the skeleton diagram, then it is the first full body occupying the picture diagram;

[0096] Second full body occupying the picture diagram:

[0097] If the right eye key point, left eye key point, right ear key point, left ear key point, nose key point, left ankle key point, and right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance from the left knee key point to the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 70% of the height of the skeleton diagram and less than or equal to 80% of the height of the skeleton diagram, then it is the second full-body image occupying the screen;

[0098] Third full-body image occupying the screen:

[0099] If the right eye key point, left eye key point, right ear key point, left ear key point, nose key point, left ankle key point, and right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance from the left knee key point to the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 60% of the height of the skeleton diagram and less than or equal to 70% of the height of the skeleton diagram, then it is the third full-body image occupying the screen;

[0100] Fourth full-body image occupying the screen:

[0101] If the right eye key point, left eye key point, right ear key point, left ear key point, nose key point, left ankle key point, and right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance from the left knee key point to the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 50% of the height of the skeleton diagram and less than or equal to 60% of the height of the skeleton diagram, then it is the fourth full-body image occupying the screen;

[0102] Fifth full-body image occupying the screen:

[0103] If the right eye key point, left eye key point, right ear key point, left ear key point, nose key point, left ankle key point, and right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance from the left knee key point to the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is less than or equal to 50% of the height of the skeleton diagram, then it is the fifth full-body image occupying the screen;

[0104] Head close-up:

[0105] If the right eye key point, left eye key point, right ear key point, left ear key point, and nose key point all exist and do not belong to the 5-point upper body diagram, then they are all head close-ups;

[0106] Make judgments in sequence. If it meets the requirements, end the judgment and label it.

[0107] In this embodiment, preferably, the comparison and recommendation module specifically: performs label matching between the model gender label, model age label, model task label, and model scene label and each scene picture in the scene picture library, and calculates the matching degree; arranges and displays the scene pictures in descending order of the matching degree;

[0108] If the gender label of the scene picture is the same as the model gender label, add x points; otherwise, add 0 points;

[0109] If the character integrity label of the scene picture is the same as the model character integrity label, add y points; otherwise, add 0 points;

[0110] If the age label of the scene picture is the same as the model age label, add z points; otherwise, add 0 points;

[0111] Compare the scene label of the scene picture with the scene label, and calculate and add m*n points according to the number m of the same ones, 5≥m≥0; x>y>z>n; each scene picture includes at least one scene label, and each model picture includes at least one scene label;

[0112] The matching degree is the sum of the above scores.

[0113] Since the device introduced in the second embodiment of the present invention is the device adopted for implementing the method in the first embodiment of the present invention, based on the method introduced in the first embodiment of the present invention, those skilled in the art can understand the specific structure and deformation of the device, so it will not be elaborated here. Any device adopted for the method in the first embodiment of the present invention belongs to the scope protected by the present invention.

[0114] Based on the same inventive concept, this application provides an electronic device embodiment corresponding to the first embodiment, as detailed in the third embodiment.

[0115] Embodiment Three

[0116] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, any implementation manner in the first embodiment can be realized.

[0117] Since the electronic device introduced in this embodiment is the device adopted for implementing the method in the first embodiment of this application, based on the method introduced in the first embodiment of this application, those skilled in the art can understand the specific implementation manner of the electronic device in this embodiment and its various forms of change. Therefore, how this electronic device implements the method in the embodiments of this application will not be introduced in detail here. Any device adopted by those skilled in the art for implementing the method in the embodiments of this application belongs to the scope protected by this application.

[0118] Based on the same inventive concept, this application provides a storage medium corresponding to Embodiment 1, details of which can be found in Embodiment 4.

[0119] Embodiment 4

[0120] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, any implementation manner in Embodiment 1 can be realized.

[0121] The technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0122] This embodiment significantly improves the accuracy of model scene graph recommendation. It not only greatly reduces the time consumption of users when screening scene graphs, but also makes the position and size of the model in the scene graph more fitting through more accurate scene graph recommendation, thus significantly improving the practicality and usability of the finally generated model graph;

[0123] Even when there is only a human integrity label in the picture provided by the user, the corresponding scene graph can still be quickly matched, greatly improving the picture selection efficiency.

[0124] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0125] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for realizing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0126] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device realizes the functions in the processFigure 1 one process or multiple processes and / or boxes Figure 1 the functions specified in one box or multiple boxes.

[0127] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one process or multiple processes and / or boxes Figure 1 one box or multiple boxes.

[0128] Although the specific embodiments of the present invention have been described above, those skilled in the art should understand that the specific embodiments we described are illustrative rather than used to limit the scope of the present invention. Equivalent modifications and variations made by those skilled in the art in accordance with the spirit of the present invention should be covered by the scope of the claims of the present invention.

Claims

1. A scene graph recommendation method based on a model graph, characterized in that: The steps include: Step 1: Set up a scene picture library, and set label information for each scene picture in the scene picture library, wherein the label information includes a gender label, an age label, a character integrity label, and a scene label; Step 2: Label the uploaded model image, including model gender label, model age label, model character integrity label and model scene label; Step 3: Match the model gender label, model age label, model character integrity label and model scene label with each scene picture in the scene picture library, and calculate the matching degree; arrange and display the scene pictures in descending order of matching degree.

2. The scene graph recommendation method based on model graph according to claim 1, characterized in that: The character completeness label includes: 5-point upper body picture, 7-point upper body picture, 9-point upper body picture, lower body picture, close-up picture of legs and feet, first full-body picture, second full-body picture, third full-body picture, fourth full-body picture, fifth full-body picture and close-up picture of head; the judgment method of the character completeness label is: Obtain a model image uploaded by a user, and obtain a skeleton image of the model image; the skeleton image includes key point information; 5-point upper body picture: If at least one of the right hip key point and the left hip key point exists, and the right hip key point or the left hip key point is less than 20 pixels from the bottom of the skeleton image, and the right knee key point and the left knee key point do not exist; then it is a 5-point upper body image; If both the left shoulder key point and the right shoulder key point exist, and the distance between the left shoulder key point and the right shoulder key point is less than 0.7 of the width of the skeleton image, and the right eye key point, the left eye key point, the right ear key point, the left ear key point and the nose key point all exist, and the distance between the right ear key point and the left ear key point is less than 1 / 3 of the width of the skeleton image; then it is a 5-point upper body image; 7 points upper body picture: If both the right hip key point and the left hip key point exist, at least one of the right knee key point and the left knee key point exists, and the right knee key point or the left knee key point is less than 20 pixels from the bottom of the skeleton image, and both the right ankle key point and the left ankle key point do not exist, then the upper body image is 7 points; 9 points upper body picture: If both the right hip key point and the left hip key point exist, both the right knee key point and the left knee key point exist, at least one of the right ankle key point and the left ankle key point exists, and the right ankle key point or the left ankle key point is less than 20 pixels from the bottom of the skeleton image, then it is a 9-point upper body image; Lower body picture: If the neck key point does not exist, the right hip key point and the left hip key point both exist, the right knee key point and the left knee key point both exist, and the right ankle key point and the left ankle key point both exist, then it is a lower body diagram; Close-up of legs and feet: If the neck key point does not exist, the right hip key point and the left hip key point do not exist, the right knee key point and the left knee key point exist, and the right ankle key point and the left ankle key point exist, then it is a close-up image of the legs and feet; First full-body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 80% of the height of the skeleton image, then it is the first full body occupancy image; Second full body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 70% of the height of the skeleton image and less than or equal to 80% of the height of the skeleton image, then it is a second full body occupancy screen image; The third full-body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 60% of the height of the skeleton image and less than or equal to 70% of the height of the skeleton image, then it is a third full body image; Fourth full body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 50% of the height of the skeleton image and less than or equal to 60% of the height of the skeleton image, then it is a fourth full body image; Fifth full body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is less than or equal to 50% of the height of the skeleton image, then it is the fifth full body occupancy image; Close-up of the head: If the right eye key point, left eye key point, right ear key point, left ear key point and nose key point all exist and are not 5-point upper body pictures, they are all head close-up pictures; Make judgments one by one. If they meet the requirements, end the judgment and add labels.

3. The scene graph recommendation method based on model graph according to claim 1, characterized in that: The step 3 specifically includes: matching the model gender label, model age label, model task label and model scene label with each scene image in the scene image library, and calculating the matching degree; arranging and displaying the scene images in descending order of matching degree; If the gender label of the scene image is the same as the model's gender label, add x points; If not, add 0 points; If the character completeness label of the scene image is the same as the character completeness label of the model, add y points; If not, add 0 points; If the age label of the scene image is the same as the model's age label, add a z-score; If not, add 0 points; Compare the scene labels of the scene graph with the scene labels, and calculate and add m*n points according to the same number m, 5≥m≥0; x>y>z>n; each scene graph includes at least one scene label, and each model graph includes at least one scene label; The matching degree is the sum of the above scores.

4. A scene graph recommendation device based on a model graph, characterized in that: include: Setting a picture library module, setting a scene picture library, and setting label information for each scene picture in the scene picture library, wherein the label information includes a gender label, an age label, a character integrity label, and a scene label; Set up a model image label module to set labels for uploaded model images, including model gender label, model age label, model character integrity label, and model scene label; The comparison recommendation module matches the model gender label, model age label, model character completeness label and model scene label with each scene picture in the scene picture library respectively, and calculates the matching degree; the scene pictures are arranged and displayed in order from high to low matching degree.

5. The scene graph recommendation device based on model graph according to claim 4, characterized in that: The character completeness label includes: 5-point upper body picture, 7-point upper body picture, 9-point upper body picture, lower body picture, close-up picture of legs and feet, first full-body picture, second full-body picture, third full-body picture, fourth full-body picture, fifth full-body picture and close-up picture of head; the judgment method of the character completeness label is: Obtain a model image uploaded by a user, and obtain a skeleton image of the model image; the skeleton image includes key point information; 5-point upper body picture: If at least one of the right hip key point and the left hip key point exists, and the right hip key point or the left hip key point is less than 20 pixels from the bottom of the skeleton image, and the right knee key point and the left knee key point do not exist; then it is a 5-point upper body image; If both the left shoulder key point and the right shoulder key point exist, and the distance between the left shoulder key point and the right shoulder key point is less than 0.7 of the width of the skeleton image, and the right eye key point, the left eye key point, the right ear key point, the left ear key point and the nose key point all exist, and the distance between the right ear key point and the left ear key point is less than 1 / 3 of the width of the skeleton image; then it is a 5-point upper body image; 7 points upper body picture: If both the right hip key point and the left hip key point exist, at least one of the right knee key point and the left knee key point exists, and the right knee key point or the left knee key point is less than 20 pixels from the bottom of the skeleton image, and both the right ankle key point and the left ankle key point do not exist, then the upper body image is 7 points; 9 points upper body picture: If both the right hip key point and the left hip key point exist, both the right knee key point and the left knee key point exist, at least one of the right ankle key point and the left ankle key point exists, and the right ankle key point or the left ankle key point is less than 20 pixels from the bottom of the skeleton image, then it is a 9-point upper body image; Lower body picture: If the neck key point does not exist, the right hip key point and the left hip key point both exist, the right knee key point and the left knee key point both exist, and the right ankle key point and the left ankle key point both exist, then it is a lower body diagram; Close-up of legs and feet: If the neck key point does not exist, the right hip key point and the left hip key point do not exist, the right knee key point and the left knee key point exist, and the right ankle key point and the left ankle key point exist, then it is a close-up image of the legs and feet; First full-body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 80% of the height of the skeleton image, then it is the first full body occupancy image; Second full body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 70% of the height of the skeleton image and less than or equal to 80% of the height of the skeleton image, then it is a second full body occupancy screen image; The third full-body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 60% of the height of the skeleton image and less than or equal to 70% of the height of the skeleton image, then it is a third full body image; Fourth full body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is greater than 50% of the height of the skeleton image and less than or equal to 60% of the height of the skeleton image, then it is a fourth full body image; Fifth full body picture: If the right eye key point, the left eye key point, the right ear key point, the left ear key point, the nose key point, the left ankle key point and the right ankle key point all exist, and the ratio of the distance between the left ankle key point and the left knee key point to the distance between the left knee key point and the left hip key point is within the set threshold, and the distance from the nose key point to the right ankle key point is less than or equal to 50% of the height of the skeleton image, then it is the fifth full body occupancy image; Close-up of the head: If the right eye key point, left eye key point, right ear key point, left ear key point and nose key point all exist and are not 5-point upper body pictures, they are all head close-up pictures; Make judgments one by one. If they meet the requirements, end the judgment and add labels.

6. The scene graph recommendation device based on model graph according to claim 4, characterized in that: The comparison and recommendation module specifically includes: matching the model gender label, model age label, model task label and model scene label with each scene picture in the scene picture library, and calculating the matching degree; arranging and displaying the scene pictures in descending order of matching degree; If the gender label of the scene image is the same as the model's gender label, add x points; If not, add 0 points; If the character completeness label of the scene image is the same as the character completeness label of the model, add y points; If not, add 0 points; If the age label of the scene image is the same as the model's age label, add a z-score; If not, add 0 points; Compare the scene labels of the scene graph with the scene labels, and calculate and add m*n points according to the same number m, 5≥m≥0; x>y>z>n; each scene graph includes at least one scene label, and each model graph includes at least one scene label; The matching degree is the sum of the above scores.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 3 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

Images