Selective 3D rendering method based on three-dimensional scene

Abstract

The invention discloses a selective 3D rendering method based on a three-dimensional scene, and particularly relates to the technical field of 3D rendering, and the method comprises the steps: determining similar users through the historical data of the users, determining the evaluation information of the users for different perfume components, determining the correlation information of different perfume component styles of the users according to the positions of perfume component labels in a knowledge graph, and carrying out the 3D rendering according to the correlation information. Before the rendering process of a user-defined perfume rendering scheme, recommendation styles of different components are determined, 3D model data of the recommendation styles are loaded in the background, and according to the user-defined perfume rendering scheme, the longest folding edge of the perfume rendering scheme at different rotation angles is discretized and analyzed, so that the 3D model data of the recommendation styles are obtained. According to the method, the rendering quality information of the perfume rendering scheme is determined, the matching degree information of the perfume rendering scheme is determined by determining the curvature characteristic change and the geometric shape change of each different mark point on the grid, and the quality and the interaction experience of user-defined perfume design can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of 3D rendering, and more particularly to a selective 3D rendering method based on a three-dimensional scene. Background Art

[0002] In the field of personalized product design and 3D visualization, users have a growing demand for customized perfume bottle design. Modern 3D rendering systems usually allow users to freely select components such as the bottle body, nozzle, bottle cap, and middle cover, and provide interactive 3D previews. However, existing systems are usually based on fixed templates or manual screening methods. The traditional 3D rendering process requires real-time loading of 3D model data of all components, which is computationally intensive and prone to delays, especially in complex rendering scenarios, affecting smoothness. In addition, the shape matching relationship between components before and after assembly in user-defined 3D design solutions may not be effectively modeled, resulting in visual or structural inconsistencies after some components are combined, affecting the overall aesthetics, and making it difficult for users to intuitively understand the final presentation effect of their customized solutions.

[0003] In order to solve the above defects, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above defects of the prior art, an embodiment of the present invention provides a selective 3D rendering method based on a three-dimensional scene to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A selective 3D rendering method based on a three-dimensional scene specifically comprises the following steps:

[0007] S1: Determine similar users through historical data of users, determine the evaluation information of users on different perfume components, and determine the correlation information of different perfume component styles of users according to the position of perfume component labels in the knowledge graph;

[0008] S2: Before the rendering process of the user-defined perfume rendering scheme, the recommended styles of different components are determined, and the 3D model data of the recommended styles are loaded in the background;

[0009] S3: according to the user-defined perfume rendering scheme, determining the rendering quality information of the perfume rendering scheme by discretizing and analyzing the longest folded edge of the perfume rendering scheme at different rotation angles;

[0010] S4: Determine the curvature feature change of each special marking point on the grid through the special marking points before and after the component combination rendering in the perfume rendering scheme, and measure the geometric shape change through the Hausdorff distance to determine the matching information of the perfume rendering scheme;

[0011] S5: Perform a comprehensive analysis on the rendering quality information and matching degree information of the perfume rendering scheme to determine the rendering performance of the user perfume rendering scheme in the three-dimensional scene.

[0012] In a preferred embodiment, determining the user's evaluation information and association information for different fragrance components includes:

[0013] Collect historical data of users' selection of perfume component styles, find other users similar to the user through collaborative filtering of historical data, and predict the target user's possible interest in unrated component styles through weighted average method based on similar users, determine the evaluation information of different perfume component styles, and represent the evaluation information of different perfume component styles through interest prediction value coefficients;

[0014] Based on the historical data of users selecting perfume component styles, determine the labels of the perfume group styles selected by users. Based on the positions of the labels of the perfume group styles in the knowledge graph, determine the degree to which users may be interested in other component styles. Determine the correlation information of different perfume component styles, and represent the correlation information of different perfume component styles through the graph connection distance coefficient.

[0015] In a preferred embodiment, the acquisition logic of the prediction value coefficient of interest is:

[0016] According to the historical data of the user's selection of the perfume component style, the score of the user's historical selection of the perfume component style is obtained, and the score of the user's historical selection of the perfume component style is marked as: , identify other users who selected the same fragrance component style as the user, and mark the scores of other users' selections of fragrance component style as: , where i=1, 2, 3, ..., I, where I is a positive integer, i is the number of the perfume component style selected by the user in history, and n=1, 2, 3, ..., N, where N is a positive integer, and n is the number of other users who selected the same perfume component style as the user;

[0017] The cosine similarity is used to measure the similarity of user behaviors, and a similarity threshold is set to filter similar users. The calculation formula is: ;in, is the similarity between the nth other user and the user;

[0018] Obtain users whose similarity with other users is greater than the similarity threshold, and record other users whose similarity with other users is greater than the similarity threshold as similar users. Calculate the user's interest prediction value coefficient for other component styles by weighted average. The calculation formula is: ;in, is the coefficient of the predicted value of interest for the kth component style.

[0019] In a preferred embodiment, the logic for obtaining the graph connection distance coefficient is:

[0020] According to the historical data of the perfume component style selected by the user, the label of the perfume component style selected by the user is obtained, and the label of the perfume component style is used as the parent node in the knowledge graph. Different perfume component styles belong to the child nodes under the parent node. The nodes are connected by edges, and the length of the edge indicates the strength of the relationship;

[0021] According to the ratings of the perfume component styles selected by the user in history, the total ratings of different parent nodes are determined. The calculation formula is: ,in, is the total score of the user at different parent nodes, is the score of the user's qth perfume component style under the mth parent node;

[0022] Determine the graph connection distance coefficient of different perfume component styles, and the calculation formula is: ;in, is the graph connection distance coefficient of the kth perfume component style, is the relationship weight between the k-th perfume component style and the parent node, and JD is the total score of the parent node of the k-th perfume component style.

[0023] In a preferred embodiment, determining recommended styles of different components includes:

[0024] The evaluation information and correlation information of different perfume component styles are comprehensively analyzed, and the recommended evaluation coefficients of different perfume component styles are determined by weighted calculation of the interest prediction value coefficient and the graph connection distance coefficient. The calculation formula of the recommended evaluation coefficient is: ;in, is the recommended evaluation coefficient, , are the proportional coefficients of the prediction value coefficient of interest and the graph connection distance coefficient, respectively. , They are both greater than 0;

[0025] Set the recommended evaluation coefficient threshold, compare the recommended evaluation coefficients of different perfume component styles with the recommended evaluation coefficient threshold. If the recommended evaluation coefficients of different perfume component styles are greater than the recommended evaluation coefficient threshold, the perfume component style will be loaded before the user customizes the perfume rendering scheme. If the recommended evaluation coefficients of different perfume component styles are less than the recommended evaluation coefficient threshold, no operation will be performed on the perfume component style.

[0026] In a preferred embodiment, determining the rendering quality information and matching degree information of the perfume rendering scheme includes:

[0027] Users form a custom perfume rendering scheme by selecting different perfume component styles. By comparing the rendering effects of the perfume components in the custom perfume rendering scheme before and after combination, the rendering quality information and matching information of the perfume rendering scheme are determined. The rendering quality information of the perfume rendering scheme is represented by the screen space error accumulation coefficient, and the matching information of the perfume rendering scheme is represented by the curvature variation coefficient and the geometric variation measurement coefficient.

[0028] In a preferred implementation, the logic for obtaining the screen space error accumulation coefficient is:

[0029] In the process of rendering the user-defined perfume rendering scheme, the longest folded edge of the perfume rendering scheme at different rotation angles is discretized and analyzed to determine the geometric error based on the longest folded edge. The geometric error of the longest folded edge at different rotation angles with the x-axis as the rotation axis is marked as: , the geometric error of the longest folded edge at different rotation angles with the y-axis as the rotation axis is marked as: , the geometric error of the longest folded edge at different rotation angles with the z-axis as the rotation axis is marked as: ,in, , J is a positive integer, , , They are the numbers of the rotation angles about the x-axis, y-axis, and z-axis respectively;

[0030] Determine the height of the perfume on the screen in the custom perfume rendering scheme, and mark the height of the perfume on the screen at different rotation angles of the x-axis, y-axis, and z-axis as follows: , as well as ;

[0031] Determine the distance between the perfume and the camera in the custom perfume rendering scheme, and mark the distance between the perfume and the camera at different rotation angles of the x-axis, y-axis, and z-axis as follows: , as well as ;

[0032] Determine the camera's field of view and mark it as FOV. Calculate the screen space error accumulation coefficient using the following formula:

[0033] ;in, is the screen space error accumulation coefficient.

[0034] In a preferred embodiment, the logic for obtaining the curvature variation coefficient is:

[0035] Determine the special marking points of each style of the perfume component, construct a discrete grid of the perfume component, determine the LBO value of the special marking point of the user-defined perfume rendering scheme before the component combination, and mark the LBO value of the special marking point of the user-defined perfume rendering scheme before the component combination as: , mark the LBO value of the special marking point of the user-defined perfume rendering scheme after the component combination as: , where g=1, 2, 3, ..., G, G is a positive integer, and g is the number of the special marking point;

[0036] Calculate the curvature variation coefficient, the calculation formula is: ;in, is the curvature variation coefficient, f(*) is the Laplace-Beltrami operator, is the L1 norm.

[0037] In a preferred embodiment, the acquisition logic of the geometric variation coefficient is:

[0038] Determine the vertex sets of the three-dimensional models of adjacent perfume components after the components are combined in the user-defined perfume rendering scheme, mark the point sets of the two adjacent components as A and B, and measure the maximum and minimum distances between the point sets of the two adjacent components by the Hausdorff distance. The calculation formula is: ; is the Hausdorff distance between two adjacent components, is the minimum distance from the points in set A to set B, is the minimum distance from the point set in set B to set A;

[0039] Obtain the Hausdorff distances of different adjacent perfume components in the user-defined perfume rendering scheme and calculate the geometric variation coefficient. The calculation formula is: ; Wherein, w=1, 2, 3, ..., L, L is a positive integer, and w is the number of different adjacent perfume components, is the geometric coefficient of variation.

[0040] In a preferred embodiment, determining the rendering performance of the user's perfume rendering scheme in the three-dimensional scene includes:

[0041] The rendering quality information and matching degree information of the perfume rendering scheme are comprehensively analyzed. The screen space error accumulation coefficient, curvature change coefficient and geometric variation measurement coefficient are weighted to construct a perfume rendering evaluation model and generate a perfume rendering evaluation coefficient. The calculation formula of the perfume rendering evaluation coefficient is: ;in, is the perfume rendering evaluation coefficient, , , They are the proportional coefficients of screen space error accumulation coefficient, curvature variation coefficient, and geometric variation measurement coefficient, respectively. , , They are both greater than 0;

[0042] Setting a perfume rendering evaluation coefficient threshold, comparing the perfume rendering evaluation coefficient of the user-defined perfume rendering scheme with the perfume rendering evaluation coefficient threshold, and generating a warning signal if the perfume rendering evaluation coefficient of the user-defined perfume rendering scheme is less than the perfume rendering evaluation coefficient threshold;

[0043] If the fragrance rendering evaluation coefficient of the user-defined fragrance rendering scheme is greater than the fragrance rendering evaluation coefficient threshold, no warning signal is generated.

[0044] Technical effects and advantages of the present invention:

[0045] The present invention optimizes the recommendation and rendering quality of user-defined perfume rendering solutions based on user historical data, knowledge graphs and 3D rendering analysis. By utilizing user historical data and knowledge graphs, similar users are determined, user preferences for different perfume components are predicted, and the correlation between perfume component styles is established. Before the user performs 3D rendering, the perfume component style that meets the user's preferences is recommended, and the corresponding 3D model is loaded in advance to improve rendering efficiency. Through discretization analysis, the geometric error of the perfume component at different rotation angles is evaluated, the longest folded edge is judged, the overall rendering quality is determined, and the degree of matching between the overall components is determined. The performance of the user-defined perfume rendering solution in the 3D scene is evaluated as a whole. The present invention helps to improve the quality and interactive experience of user-defined perfume design, while reducing the rendering calculation burden and improving system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0047] Figure 1 The present invention is a flowchart of a selective 3D rendering method based on a three-dimensional scene. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Example 1

[0050] Figure 1 The flowchart of a selective 3D rendering method based on a three-dimensional scene is shown in FIG. 1 , which specifically includes the following steps:

[0051] S1: Determine similar users through historical data of users, determine the evaluation information of users on different perfume components, and determine the correlation information of different perfume component styles of users according to the position of perfume component labels in the knowledge graph;

[0052] S2: Before the rendering process of the user-defined perfume rendering scheme, the recommended styles of different components are determined, and the 3D model data of the recommended styles are loaded in the background;

[0053] S3: according to the user-defined perfume rendering scheme, determining the rendering quality information of the perfume rendering scheme by discretizing and analyzing the longest folded edge of the perfume rendering scheme at different rotation angles;

[0054] S4: Determine the curvature feature change of each special marking point on the grid through the special marking points before and after the component combination rendering in the perfume rendering scheme, and measure the geometric shape change through the Hausdorff distance to determine the matching information of the perfume rendering scheme;

[0055] S5: Perform a comprehensive analysis on the rendering quality information and matching degree information of the perfume rendering scheme to determine the rendering performance of the user perfume rendering scheme in the three-dimensional scene.

[0056] In the design process of a custom perfume scheme, perfume is also composed of multiple components, including the bottle body, nozzle, bottle cap and middle cover. Different component types include multiple styles. The user-defined perfume rendering scheme is determined by combining different styles in different component types.

[0057] Specifically, users select different component styles according to their needs. The rendering content and style of each component will be different. Different components may have different levels of detail or complexity, so only the part currently selected by the user will be loaded and displayed during rendering;

[0058] For different style selections, users can dynamically adjust their material and color properties instead of statically rendering every possible combination, including dynamic adjustment rendering of materials and colors, that is, rendering the corresponding appearance according to the user's selection.

[0059] Collect historical data of users' selection of perfume component styles, find other users similar to the user through collaborative filtering of historical data, and predict the degree to which the target user may be interested in unrated component styles through weighted average method based on similar users, determine the evaluation information of different perfume component styles, and represent the evaluation information of different perfume component styles through interest prediction value coefficients, determine the label of the perfume group style selected by the user based on the historical data of the user's selection of perfume component styles, determine the degree to which the user may be interested in other component styles based on the position of the label of the perfume group style in the knowledge graph, determine the correlation information of different perfume component styles, and represent the correlation information of different perfume component styles through the graph connection distance coefficient.

[0060] The logic for obtaining the coefficient of interest prediction value is: according to the historical data of the user selecting the perfume component style, obtain the score of the user's historical selection of the perfume component style, and mark the score of the user's historical selection of the perfume component style as: , identify other users who selected the same fragrance component style as the user, and mark the scores of other users' selections of fragrance component style as: , where i=1, 2, 3, ..., I, where I is a positive integer, i is the number of the perfume component style selected by the user in history, n=1, 2, 3, ..., N, where N is a positive integer, n is the number of other users who selected the same perfume component style as the user;

[0061] It should be noted that the common point between the user and other users who have selected the same perfume component style as the user is that they have all used the same perfume component style in their historical usage records. It should be noted that other users who have selected the same perfume component style as the user have not only used the same perfume component style as the user.

[0062] The cosine similarity is used to measure the similarity of user behaviors, and a similarity threshold is set to filter similar users. The calculation formula is: ;in, is the similarity between the nth other user and the user;

[0063] Obtain users whose similarity with other users is greater than the similarity threshold, and record other users whose similarity with other users is greater than the similarity threshold as similar users. Calculate the user's interest prediction value coefficient for other component styles by weighted average. The calculation formula is: ;in, is the coefficient of the predicted value of interest for the kth component style.

[0064] It can be seen from the formula that the larger the interest prediction value coefficient of a component style is, the more likely the user is to be interested in the component style. The component style can be loaded before the user customizes the perfume rendering scheme to reduce the burden of rendering calculation.

[0065] The logic for obtaining the graph connection distance coefficient is as follows: according to the historical data of the user selecting the perfume component style, obtain the label of the user's historical selection of the perfume component style, use the label of the perfume component style as the parent node in the knowledge graph, and different perfume component styles belong to the child nodes under the parent node. The nodes are connected by edges, and the length of the edge indicates the strength of the relationship;

[0066] It should be noted that the labels of the perfume component style include fragrance labels such as floral, fruity, and woody, design style labels such as modern simplicity, classic, and luxury, and functional labels such as durability, leak-proof design, and easy cleaning. Different parent nodes may have the same child nodes, and the edges between parent nodes and child nodes are represented by relationship weights.

[0067] According to the ratings of the perfume component styles selected by the user in history, the total ratings of different parent nodes are determined. The calculation formula is: ,in, is the total score of the user at different parent nodes, is the score of the user's qth perfume component style under the mth parent node;

[0068] It should be noted that the total scores of different nodes represent the user's preferences for different tags. The larger the total scores of different nodes, the more interested the user is in the perfume component style under the tag.

[0069] Determine the graph connection distance coefficient of different perfume component styles, and the calculation formula is: ;in, is the graph connection distance coefficient of the kth perfume component style, is the relationship weight between the k-th perfume component style and the parent node, and JD is the total score of the parent node of the k-th perfume component style.

[0070] It should be noted that the larger the graph connection distance coefficient is, the more likely the perfume component style is to be in line with the user's choice. The component style can be loaded before the user customizes the perfume rendering scheme to reduce the burden of rendering calculations.

[0071] The evaluation information and correlation information of different perfume component styles are comprehensively analyzed, and the recommended evaluation coefficients of different perfume component styles are determined by weighted calculation of the interest prediction value coefficient and the graph connection distance coefficient. The calculation formula of the recommended evaluation coefficient is: ;in, is the recommended evaluation coefficient, , are the proportional coefficients of the prediction value coefficient of interest and the graph connection distance coefficient, respectively. , Both are greater than 0.

[0072] Set the recommended evaluation coefficient threshold, and compare the recommended evaluation coefficients of different perfume component styles with the recommended evaluation coefficient threshold. If the recommended evaluation coefficients of different perfume component styles are greater than the recommended evaluation coefficient threshold, the perfume component style will be loaded before the user customizes the perfume rendering scheme, indicating that the perfume component style may be applied by the user, and the 3D model data of the perfume component style needs to be loaded in advance. If the recommended evaluation coefficients of different perfume component styles are less than the recommended evaluation coefficient threshold, no operation will be performed on the perfume component style.

[0073] Users form a custom perfume rendering scheme by selecting different perfume component styles. By comparing the rendering effects of the perfume components in the custom perfume rendering scheme before and after combination, the rendering quality information and matching information of the perfume rendering scheme are determined. The rendering quality information of the perfume rendering scheme is represented by the screen space error accumulation coefficient, and the matching information of the perfume rendering scheme is represented by the curvature variation coefficient and the geometric variation measurement coefficient.

[0074] The logic for obtaining the screen space error accumulation coefficient is as follows: in the process of rendering the user-defined perfume rendering scheme, the longest folded edge of the perfume rendering scheme at different rotation angles is analyzed by discretization, and the geometric error based on the longest folded edge is determined, and the geometric error of the longest folded edge at different rotation angles with the x-axis as the rotation axis is marked as: , the geometric error of the longest folded edge at different rotation angles with the y-axis as the rotation axis is marked as: , the geometric error of the longest folded edge at different rotation angles with the z-axis as the rotation axis is marked as: ,in, , J is a positive integer, , , They are the numbers of the rotation angles about the x-axis, y-axis, and z-axis respectively;

[0075] It should be noted that when folding edges, a certain error will occur after the two endpoints are merged. This error is usually evaluated by calculating the displacement between the vertices before and after the merger. The size of the error will affect the accuracy of the simplified model. During the simplification process, the longest edge is usually selected for folding, because the folding of long edges usually produces a large geometric error, which affects the appearance of the model;

[0076] In 3D computing, the geometric information of an object is usually represented by a set of vertices, edges, and faces. In order to improve rendering performance, it is often necessary to simplify these models and reduce the number of vertices or faces. The simplified model is not completely equal to the original model and may lose some details. This distortion or deviation is the geometric error.

[0077] Determining the geometric error of the longest folded edge at different rotation angles with the x, y, and z axes as the rotation axes can optimize the accuracy of 3D model simplification or model presentation while ensuring the visual quality at different viewing angles. By discretizing the rotation angles, the angle of each rotation axis is discretized to ensure that the amount of calculation can be reduced and a reasonable result can be provided in a short time. The discretized rotation angles are set by professional staff.

[0078] Determine the height of the perfume on the screen in the custom perfume rendering scheme, and mark the height of the perfume on the screen at different rotation angles of the x-axis, y-axis, and z-axis as follows: , as well as ;

[0079] It should be noted that the height of the perfume on the screen refers to the projection height of the object after rendering. In a three-dimensional scene, the perfume is converted to the screen coordinate system through the projection matrix and then mapped to the pixel area on the screen. When the perfume is rotated, the projection height of the perfume on the screen may change, especially when the rotation angle of the perfume causes its shape and outline to change.

[0080] Determine the distance between the perfume and the camera in the custom perfume rendering scheme, and mark the distance between the perfume and the camera at different rotation angles of the x-axis, y-axis, and z-axis as follows: , as well as ;

[0081] It should be noted that the distance between the camera and the perfume refers to the straight-line distance from the object to the center of the camera, which is a key factor affecting the size, clarity and perspective effect of the perfume displayed on the screen. When the perfume is rotated, the perfume may be toward or away from the camera, so the distance between the camera and the perfume will change.

[0082] Determine the camera's field of view and mark it as FOV. Calculate the screen space error accumulation coefficient using the following formula:

[0083] ;in, is the screen space error accumulation coefficient.

[0084] It can be seen from the formula that the larger the screen space error accumulation coefficient is, the larger the error is in the rendering process of the user-defined perfume rendering scheme, which means the rendering effect is worse. Therefore, the user-defined perfume rendering scheme performs poorly in three-dimensional scenes, and it may be necessary to improve the rendering accuracy and effect through appropriate optimization and adjustment.

[0085] The acquisition logic of the curvature variation coefficient is: determine the special marking points of each style of the perfume component, construct a discrete grid of the perfume component, determine the LBO value of the special marking point of the user-defined perfume rendering scheme before the component combination, and mark the LBO value of the special marking point of the user-defined perfume rendering scheme before the component combination as: , mark the LBO value of the special marking point of the user-defined perfume rendering scheme after the component combination as: , where g=1, 2, 3, ..., G, G is a positive integer, and g is the number of the special marking point;

[0086] It should be noted that in the 3D model of the perfume component, points with important geometric or visual features are selected as special marking points, such as connection points, boundary points, points with large curvature changes, etc. Special marking points are set by professional staff. Special marking points may be deformed and have alignment errors when different components are combined, so their LBO values ​​can be used as important parameters to measure geometric changes.

[0087] Calculate the curvature variation coefficient, the calculation formula is: ;in, is the curvature variation coefficient, f(*) is the Laplace-Beltrami operator, is the L1 norm.

[0088] It can be seen from the formula that the larger the curvature variation coefficient is, the more significant the local curvature changes due to the splicing of components, and the surface in the transition area between components changes suddenly, which may affect the rendering quality. Compared with the original components, the surface structure of the combined form changes significantly, resulting in an increase in the LBO difference.

[0089] The logic for obtaining the geometric variation coefficient is as follows: determine the vertex set of the three-dimensional model of adjacent perfume components after the components are combined in the user-defined perfume rendering scheme, mark the point sets of two adjacent components as A and B, and measure the maximum and minimum distance between the point sets of two adjacent components by Hausdorff distance. The calculation formula is: ; is the Hausdorff distance between two adjacent components, is the minimum distance from the points in set A to set B, is the minimum distance from the point set in set B to set A;

[0090] It should be noted that adjacent perfume components include the spray head and the middle cover, the bottle body and the bottle cap, the spray head and the bottle body, the spray head and the bottle cap, the bottle body and the middle cover, the bottle cap and the middle cover, etc., which are specifically determined by the user-defined perfume rendering scheme.

[0091] Obtain the Hausdorff distances of different adjacent perfume components in the user-defined perfume rendering scheme and calculate the geometric variation coefficient. The calculation formula is: ; Wherein, w=1, 2, 3, ..., L, L is a positive integer, and w is the number of different adjacent perfume components, is the geometric coefficient of variation.

[0092] It can be seen from the formula that the larger the geometric variation coefficient, the lower the matching degree between components, the larger the assembly error, or the more serious the rendering deformation. The Hausdorff distance can be reduced by optimizing the component shape, improving the alignment accuracy, adjusting the rotation angle, and optimizing the rendering parameters, thereby improving the 3D assembly accuracy and rendering quality of the perfume bottle component.

[0093] The rendering quality information and matching degree information of the perfume rendering scheme are comprehensively analyzed. The screen space error accumulation coefficient, curvature change coefficient and geometric variation measurement coefficient are weighted to construct a perfume rendering evaluation model and generate a perfume rendering evaluation coefficient. The calculation formula of the perfume rendering evaluation coefficient is: ;in, is the perfume rendering evaluation coefficient, , , They are the proportional coefficients of screen space error accumulation coefficient, curvature variation coefficient, and geometric variation measurement coefficient, respectively. , , Both are greater than 0.

[0094] It can be seen from the formula that the larger the screen space error accumulation coefficient, the curvature change coefficient and the geometric variation measurement coefficient, the smaller the perfume rendering evaluation coefficient, which means that the user-defined perfume rendering scheme has a poor rendering effect in the three-dimensional scene. Conversely, the smaller the screen space error accumulation coefficient, the curvature change coefficient and the geometric variation measurement coefficient, the larger the perfume rendering evaluation coefficient, which means that the user-defined perfume rendering scheme has a better rendering effect in the three-dimensional scene.

[0095] A threshold value for the perfume rendering evaluation coefficient is set, and the perfume rendering evaluation coefficient of the user-defined perfume rendering scheme is compared with the perfume rendering evaluation coefficient threshold value. If the perfume rendering evaluation coefficient of the user-defined perfume rendering scheme is less than the perfume rendering evaluation coefficient threshold value, a warning signal is generated, indicating that the rendering effect of the customized perfume rendering scheme in the three-dimensional scene is poor. If the perfume rendering evaluation coefficient of the user-defined perfume rendering scheme is greater than the perfume rendering evaluation coefficient threshold value, no warning signal is generated, indicating that the rendering effect of the customized perfume rendering scheme in the three-dimensional scene is good, indicating that the user-defined perfume rendering scheme can display the actual perfume situation.

[0096] The present invention optimizes the recommendation and rendering quality of user-defined perfume rendering solutions based on user historical data, knowledge graphs and 3D rendering analysis. By utilizing user historical data and knowledge graphs, similar users are determined, user preferences for different perfume components are predicted, and the correlation between perfume component styles is established. Before the user performs 3D rendering, the perfume component style that meets the user's preferences is recommended, and the corresponding 3D model is loaded in advance to improve rendering efficiency. Through discretization analysis, the geometric error of the perfume component at different rotation angles is evaluated, the longest folded edge is judged, the overall rendering quality is determined, and the degree of matching between the overall components is determined. The performance of the user-defined perfume rendering solution in the 3D scene is evaluated as a whole. The present invention helps to improve the quality and interactive experience of user-defined perfume design, while reducing the rendering calculation burden and improving system performance.

[0097] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0098] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0099] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0100] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0101] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0102] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0103] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A selective 3D rendering method based on a three-dimensional scene, characterized in that: The specific steps include: S1: Determine similar users through historical data of users, determine the evaluation information of users on different perfume components, and determine the correlation information of different perfume component styles of users according to the position of perfume component labels in the knowledge graph; S2: Before the rendering process of the user-defined perfume rendering scheme, the recommended styles of different components are determined, and the 3D model data of the recommended styles are loaded in the background; S3: according to the user-defined perfume rendering scheme, determining the rendering quality information of the perfume rendering scheme by discretizing and analyzing the longest folded edge of the perfume rendering scheme at different rotation angles; S4: Determine the curvature feature change of each special marking point on the grid through the special marking points before and after the component combination rendering in the perfume rendering scheme, and measure the geometric shape change through the Hausdorff distance to determine the matching information of the perfume rendering scheme; S5: Perform a comprehensive analysis on the rendering quality information and matching degree information of the perfume rendering scheme to determine the rendering performance of the user perfume rendering scheme in the three-dimensional scene.

2. The selective 3D rendering method based on a three-dimensional scene according to claim 1, characterized in that: Determine user evaluation information and relevance information for different fragrance components, including: Collect historical data of users' selection of perfume component styles, find other users similar to the user through collaborative filtering of historical data, and predict the target user's possible interest in unrated component styles through weighted average method based on similar users, determine the evaluation information of different perfume component styles, and represent the evaluation information of different perfume component styles through interest prediction value coefficients; Based on the historical data of users selecting perfume component styles, determine the labels of the perfume group styles selected by users. Based on the positions of the labels of the perfume group styles in the knowledge graph, determine the degree to which users may be interested in other component styles. Determine the correlation information of different perfume component styles, and represent the correlation information of different perfume component styles through the graph connection distance coefficient.

3. The selective 3D rendering method based on a three-dimensional scene according to claim 2, characterized in that: The acquisition logic of the predicted value coefficient of interest is: According to the historical data of the user's selection of the perfume component style, the score of the user's historical selection of the perfume component style is obtained, and the score of the user's historical selection of the perfume component style is marked as: , identify other users who selected the same fragrance component style as the user, and mark the scores of other users' selections of fragrance component style as: , where i=1, 2, 3, ..., I, where I is a positive integer, i is the number of the perfume component style selected by the user in history, and n=1, 2, 3, ..., N, where N is a positive integer, and n is the number of other users who selected the same perfume component style as the user; The cosine similarity is used to measure the similarity of user behaviors, and a similarity threshold is set to filter similar users. The calculation formula is: ;in, is the similarity between the nth other user and the user; Obtain users whose similarity with other users is greater than the similarity threshold, and record other users whose similarity with other users is greater than the similarity threshold as similar users. Calculate the user's interest prediction value coefficient for other component styles by weighted average. The calculation formula is: ;in, is the coefficient of the predicted value of interest for the kth component style.

4. The selective 3D rendering method based on a three-dimensional scene according to claim 3, characterized in that: The logic for obtaining the graph connection distance coefficient is: According to the historical data of the perfume component style selected by the user, the label of the perfume component style selected by the user is obtained, and the label of the perfume component style is used as the parent node in the knowledge graph. Different perfume component styles belong to the child nodes under the parent node. The nodes are connected by edges, and the length of the edge indicates the strength of the relationship; According to the ratings of the perfume component styles selected by the user in history, the total ratings of different parent nodes are determined. The calculation formula is: ,in, is the total score of the user in different parent nodes, is the score of the user's qth perfume component style under the mth parent node; Determine the graph connection distance coefficient of different perfume component styles, and the calculation formula is: ;in, is the graph connection distance coefficient of the kth perfume component style, is the relationship weight between the k-th perfume component style and the parent node, and JD is the total score of the parent node of the k-th perfume component style.

5. The selective 3D rendering method based on a three-dimensional scene according to claim 4, characterized in that: Identify recommended styles for different components, including: The evaluation information and correlation information of different perfume component styles are comprehensively analyzed, and the recommended evaluation coefficients of different perfume component styles are determined by weighted calculation of the interest prediction value coefficient and the graph connection distance coefficient. The calculation formula of the recommended evaluation coefficient is: ;in, is the recommended evaluation coefficient, , are the proportional coefficients of the prediction value coefficient of interest and the graph connection distance coefficient, respectively. , They are both greater than 0; Set the recommended evaluation coefficient threshold, compare the recommended evaluation coefficients of different perfume component styles with the recommended evaluation coefficient threshold. If the recommended evaluation coefficients of different perfume component styles are greater than the recommended evaluation coefficient threshold, the perfume component style will be loaded before the user customizes the perfume rendering scheme. If the recommended evaluation coefficients of different perfume component styles are less than the recommended evaluation coefficient threshold, no operation will be performed on the perfume component style.

6. The selective 3D rendering method based on a three-dimensional scene according to claim 5, characterized in that: Determine the rendering quality information and matching information of the perfume rendering scheme, including: Users form a custom perfume rendering scheme by selecting different perfume component styles. By comparing the rendering effects of the perfume components in the custom perfume rendering scheme before and after combination, the rendering quality information and matching information of the perfume rendering scheme are determined. The rendering quality information of the perfume rendering scheme is represented by the screen space error accumulation coefficient, and the matching information of the perfume rendering scheme is represented by the curvature variation coefficient and the geometric variation measurement coefficient.

7. The selective 3D rendering method based on a three-dimensional scene according to claim 6, characterized in that: The logic for obtaining the screen space error accumulation coefficient is: In the process of rendering the user-defined perfume rendering scheme, the longest folded edge of the perfume rendering scheme at different rotation angles is discretized and analyzed to determine the geometric error based on the longest folded edge. The geometric error of the longest folded edge at different rotation angles with the x-axis as the rotation axis is marked as: , the geometric error of the longest folded edge at different rotation angles with the y-axis as the rotation axis is marked as: , the geometric error of the longest folded edge at different rotation angles with the z-axis as the rotation axis is marked as: ,in, , J is a positive integer, , , They are the numbers of the rotation angles about the x-axis, y-axis, and z-axis respectively; Determine the height of the perfume on the screen in the custom perfume rendering scheme, and mark the height of the perfume on the screen at different rotation angles of the x-axis, y-axis, and z-axis as follows: , as well as ; Determine the distance between the perfume and the camera in the custom perfume rendering scheme, and mark the distance between the perfume and the camera at different rotation angles of the x-axis, y-axis, and z-axis as follows: , as well as ; Determine the camera's field of view and mark it as FOV. Calculate the screen space error accumulation coefficient using the following formula: ;in, is the screen space error accumulation coefficient.

8. The selective 3D rendering method based on a three-dimensional scene according to claim 7, characterized in that: The acquisition logic of the curvature variation coefficient is: Determine the special marking points of each style of the perfume component, construct a discrete grid of the perfume component, determine the LBO value of the special marking point of the user-defined perfume rendering scheme before the component combination, and mark the LBO value of the special marking point of the user-defined perfume rendering scheme before the component combination as: , mark the LBO value of the special marking point of the user-defined perfume rendering scheme after the component combination as: , where g=1, 2, 3, ..., G, G is a positive integer, and g is the number of the special marking point; Calculate the curvature variation coefficient, the calculation formula is: ;in, is the curvature variation coefficient, f(*) is the Laplace-Beltrami operator, is the L1 norm.

9. The selective 3D rendering method based on a three-dimensional scene according to claim 8, characterized in that: The acquisition logic of the geometric variation coefficient is: Determine the vertex sets of the three-dimensional models of adjacent perfume components after the components are combined in the user-defined perfume rendering scheme, mark the point sets of the two adjacent components as A and B, and measure the maximum and minimum distances between the point sets of the two adjacent components by the Hausdorff distance. The calculation formula is: ; is the Hausdorff distance between two adjacent components, is the minimum distance from the points in set A to set B, is the minimum distance from the point set in set B to set A; Obtain the Hausdorff distances of different adjacent perfume components in the user-defined perfume rendering scheme and calculate the geometric variation coefficient. The calculation formula is: ; Wherein, w=1, 2, 3, ..., L, L is a positive integer, and w is the number of different adjacent perfume components, is the geometric coefficient of variation.

10. The selective 3D rendering method based on a three-dimensional scene according to claim 9, characterized in that: Determine how the user's perfume rendering solution will render in the 3D scene, including: The rendering quality information and matching degree information of the perfume rendering scheme are comprehensively analyzed. The screen space error accumulation coefficient, curvature change coefficient and geometric variation measurement coefficient are weighted to construct a perfume rendering evaluation model and generate a perfume rendering evaluation coefficient. The calculation formula of the perfume rendering evaluation coefficient is: ;in, is the perfume rendering evaluation coefficient, , , They are the proportional coefficients of screen space error accumulation coefficient, curvature variation coefficient, and geometric variation measurement coefficient, respectively. , , They are both greater than 0; Setting a perfume rendering evaluation coefficient threshold, comparing the perfume rendering evaluation coefficient of the user-defined perfume rendering scheme with the perfume rendering evaluation coefficient threshold, and generating a warning signal if the perfume rendering evaluation coefficient of the user-defined perfume rendering scheme is less than the perfume rendering evaluation coefficient threshold; If the fragrance rendering evaluation coefficient of the user-defined fragrance rendering scheme is greater than the fragrance rendering evaluation coefficient threshold, no warning signal is generated.

Images