An interactive model display method and system

Abstract

The present invention relates to the technical field of model display, and particularly to an interactive model display method and system, including: the model construction unit constructs a model according to a user's instruction, obtains the contour features of each sub-model, compares the contour features with corresponding preset contour features, rates the sub-models according to the comparison results, and determines whether the construction of the model meets the standard according to the rating results. When it is determined that the construction of the model does not meet the standard, the construction parameters of the model are adjusted, including controlling the correction unit to update the coordinates of the corresponding sub-model, or correcting the learning rate of model training during the model display process, which effectively improves the accuracy of model generation and further improves the generation efficiency of the model.

Description

Technical Field

[0001] The present invention relates to the technical field of model display, and particularly to an interactive model display method and system. Background Art

[0002] With the mobility of data platforms, a large amount of data is seeking mobile display, especially 3D model data. The development of technology has enabled more and more fields to apply three-dimensional modeling technology, bringing many conveniences to people's lives. Especially in the e-commerce field, using three-dimensional models to display products instead of traditional pictures gives users a clearer and more intuitive understanding and makes it easier for users to understand the detailed features of products.

[0003] Chinese Patent Publication No.: CN108629850B discloses a method for realizing mobile display interaction of 3D models, including: First, obtain 3D model data from a data center, perform decimation optimization on the 3D model, and put it into a specified folder; Second, check whether the current environment is correct, and set the environment if it is incorrect; Third, import it into different layer areas according to different categories of 3D models; Fourth, create a Point Light in each area and name it according to which space is entered; Fifth, perform material adaptation, generate a normal map for the 3D model, and put it into the normal map channel; Sixth, determine the texture specification and texture the 3D model in the imported file; Seventh, upload the model to the background page, introduce U3D, select the 3D model to be exported and uploaded, call out the corresponding upload options, and upload the 3D model as a Prefabs file; It can be seen that the above technical solution has the following problems: It does not consider determining whether the generation of the model is qualified according to the rating results of each sub-model in the generated model, which affects the accuracy of the generated model and thus affects the generation efficiency of the model. Summary of the Invention

[0004] Therefore, the present invention provides an interactive model display method and system to overcome the problem in the prior art that it does not consider determining whether the generation of the model is qualified according to the rating results of each sub-model in the generated model, which affects the accuracy of the generated model and thus affects the generation efficiency of the model.

[0005] On the one hand, the present invention provides an interactive model display method, including:

[0006] Input information through an interaction unit so that the interaction unit outputs a display instruction and an angle instruction, and an instruction transmission unit conveys the display instruction and the angle instruction to a model construction unit;

[0007] The model construction unit obtains corresponding model information from the data storage unit based on the received display instruction and completes the construction of the model based on the model information. At the same time, it rotates the constructed model to the corresponding angle based on the angle instruction. Among them, a single model information includes several sub-model information. The model construction unit constructs the corresponding sub-models in sequence based on each sub-model information and constrains each sub-model under the corresponding coordinates to complete the construction of the model;

[0008] Use the image acquisition unit to obtain the image information of the model from the corresponding angle, and obtain the contour features of each sub-model at the current angle based on the image information;

[0009] Use the analysis unit to rate each sub-model in sequence based on the contour features of each sub-model, and determine whether the construction of the model meets the standard based on the number of sub-models at each level after rating. When the analysis unit determines that the construction of the model does not meet the standard, it controls the correction unit to update the coordinates of the corresponding sub-model, or corrects the learning rate of model training during the model display process;

[0010] When the analysis unit determines that the construction of the model meets the standard, it controls the rendering unit to start rendering the model, and outputs the rendered model to the display unit after rendering.

[0011] Further, the analysis unit determines whether the construction of a single sub-model is qualified based on the coincidence degree between the contour features of each sub-model determined by the model construction unit at the corresponding angle and the preset contour features corresponding to a single sub-model, including:

[0012] If the coincidence degree is greater than the preset coincidence degree, it is determined that the construction of the sub-model is qualified, and the sub-model is recorded as a first-level sub-model;

[0013] If the coincidence degree is less than or equal to the preset coincidence degree, it is determined that the construction of the sub-model is unqualified, and the single sub-model is rated based on the dispersion amount of the coincidence abnormal area between the contour feature and the preset contour feature.

[0014] Further, the analysis unit rates a single sub-model based on the dispersion amount of the coincidence abnormal area between the contour feature and the preset contour feature, including:

[0015] For a single sub-model, the analysis unit marks the coincidence abnormal area between the contour feature and the preset contour feature as an abnormal segment to obtain the shortest distance between each abnormal segment, and determines the ratio of the average value of each shortest distance to the average value of the lengths of each abnormal segment as the dispersion amount for the sub-model;

[0016] If the dispersion amount is less than or equal to the first preset dispersion amount, the single sub-model is recorded as a second-level sub-model;

[0017] If the dispersion amount is less than or equal to the second preset dispersion amount and greater than the first preset dispersion amount, re-rate a single sub-model based on the number of sub-models;

[0018] If the dispersion amount is greater than the second preset dispersion amount, mark a single sub-model as a third-level sub-model.

[0019] Further, the analysis unit re-rates a single sub-model based on the total number of sub-models, including:

[0020] If the total number is greater than the preset total number, mark a single sub-model as a first-level sub-model, and update the current profile feature of the single sub-model to the preset profile feature for the current instruction;

[0021] If the total number is less than or equal to the preset total number, mark a single sub-model as a third-level sub-model.

[0022] Further, when the analysis unit completes the rating of each sub-model, determine whether the construction of the model meets the standard according to the level model with the largest quantity proportion in each level of sub-models, including:

[0023] If the level model with the largest quantity proportion is a first-level sub-model, determine that the construction of the model meets the standard and render the model;

[0024] If the level model with the largest quantity proportion is a second-level sub-model, determine that the construction of the model does not meet the standard and correct the vertex coordinates of each second-level sub-model;

[0025] If the level model with the largest quantity proportion is a third-level sub-model, determine that the construction of the model does not meet the standard and adjust the learning rate to the corresponding value according to the quantity proportion of the third-level sub-models.

[0026] Further, the correction unit is used to correct the vertex coordinates of each second-level sub-model;

[0027] For a single second-level sub-model, the correction unit is used to determine the offset according to the dispersion difference between the first preset dispersion amount and the dispersion amount, where

[0028] The offset determined based on the dispersion difference is proportional to the dispersion difference.

[0029] Further, when the correction unit completes the correction of the vertex coordinates of each secondary sub-model, it re-determines whether the construction of each sub-model is qualified based on the coincidence degree between the obtained contour features of each secondary sub-model and the preset contour features, including: if the coincidence degree is greater than the preset coincidence degree, it determines that the construction of the sub-model is qualified and records the sub-model as a primary sub-model; if the coincidence degree is less than or equal to the preset coincidence degree, it determines that the construction of the sub-model is unqualified and adjusts the rotation angle of a single secondary sub-model to the corresponding value according to the difference between the preset coincidence degree and the coincidence degree;

[0030] The correction unit is used to determine the rotation angle for a single secondary sub-model based on the coincidence difference between the preset coincidence degree and the coincidence degree, where,

[0031] The determined rotation angle is proportional to the coincidence difference.

[0032] Further, the correction unit determines the learning rate based on the proportion of the number of tertiary sub-models, where:

[0033] The learning rate determined based on the proportion of the number of tertiary sub-models is inversely proportional to the proportion of the number of tertiary sub-models.

[0034] On the other hand, the present invention also provides an interactive model display system using the above method, including:

[0035] An interaction unit, which is used to output a display instruction and an angle instruction based on the input information;

[0036] An instruction transmission unit, which is connected to the interaction unit and is used to transmit the display instruction and the angle instruction;

[0037] A data storage unit, which is used to store a number of model information, and a single model information includes a number of sub-model information;

[0038] A model construction unit, which is respectively connected to the instruction transmission unit and the data storage unit, and is used to complete the construction of the model based on the received display instruction and the obtained corresponding model information, and rotate the constructed model to the corresponding angle based on the angle instruction;

[0039] An image acquisition unit, which is connected to the model construction unit and is used to acquire the image information of the model and obtain the contour features of each sub-model based on the image information;

[0040] An analysis unit, which is connected to the image acquisition unit and is used to rate each sub-model in turn based on the contour features of each sub-model, and determine whether the construction of the model meets the standard based on the number of sub-models of each level after rating. When the analysis unit determines that the construction of the model does not meet the standard, it determines the corresponding instruction;

[0041] A correction unit, which is respectively connected to the model construction unit and the analysis unit, is used to execute the instructions of the analysis unit, including updating the coordinates of the corresponding sub-model, or correcting the learning rate of model training during the model display process;

[0042] A rendering unit, which is respectively connected to the analysis unit and the model construction unit, is used to render the model when it is determined that the construction of the model meets the standards;

[0043] A display unit, which is connected to the rendering unit, is used to display the rendered model.

[0044] Compared with the prior art, the beneficial effects of the present invention are that the model construction unit constructs the model according to the user's instructions, obtains the contour features of each sub-model, compares the contour features with the corresponding preset contour features, rates the sub-models according to the comparison results, and determines whether the construction of the model meets the standards according to the rating results, so as to adjust the construction parameters of the model when it is determined that the construction of the model does not meet the standards, including controlling the correction unit to update the coordinates of the corresponding sub-model, or correcting the learning rate of model training during the model display process, which effectively improves the accuracy of model generation and further improves the generation efficiency of the model.

[0045] Further, for rating a single sub-model, when the overlap degree between the obtained contour features and the preset contour features is less than or equal to the preset overlap degree, there may be an abnormality in the construction of the sub-model. The dispersion amount is used to rate a single sub-model, and the dispersion amount represents the distribution of the overlapping abnormal areas. When the dispersion amount is less than or equal to the first preset dispersion amount, the overlapping abnormal areas are concentratedly distributed. In this case, there is a situation where the position of a single sub-model is shifted, resulting in partial occlusion of the sub-model. At this time, the single occluded sub-model is marked as a secondary sub-model; when the dispersion amount is greater than the second preset dispersion amount, the overlapping abnormal areas are dispersedly distributed. In this case, there are multiple overlapping abnormal areas in the sub-model. Since the learning rate is too large, the distance spanned by the model when updating the weights each time is too large, exceeding the range within which the model can correctly adjust to adapt to the data and the target contour, resulting in the model being continuously strengthened in the wrong area and being difficult to correct the error and move in the correct direction, thus causing abnormal phenomena of unstable and inaccurate contour fitting. At this time, the sub-model with contour features having multiple abnormalities is marked as a tertiary model; when the dispersion amount is less than or equal to the second preset dispersion amount and greater than the first preset dispersion amount, the overlapping abnormal areas are evenly distributed. At this time, the total number of sub-models is determined. When the total number is greater than the preset total number, the occlusion situation in each sub-model is complex. To ensure the flexibility of model display and the tolerance for each sub-model, although the overlap degree of the contours is low, to ensure the integrity of the model architecture, its existence is accepted, and a single sub-model is determined as a primary model. By accurately rating each sub-model in turn to numerically determine the generation situation of the model, the generation accuracy of the model is effectively improved.

[0046] Further, it is determined whether the construction of the model meets the standard according to the proportion of sub-models of each level. When the level model with the largest number proportion is the secondary sub-model, there are a large number of position shifts of sub-models, resulting in partial occlusion of the sub-models. The vertex coordinates of each secondary sub-model are corrected to ensure the reasonable placement position and hierarchical structure of each sub-model, effectively avoiding mutual occlusion; when the level model with the largest number proportion is the tertiary sub-model, there are unstable and inaccurate contour fittings of a large number of sub-models. At this time, the learning rate is adjusted to improve the reliability of sub-model contour display, further improving the generation efficiency of the model. Description of the Drawings

[0047] Figure 1 It is a flowchart of the steps of the interactive model display method according to an embodiment of the present invention;

[0048] Figure 2 It is a block diagram of the modules of the interactive model display system according to an embodiment of the present invention. Detailed Embodiments

[0049] In order to make the objectives and advantages of the present invention more clearly understood, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0050] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0051] It should be noted that in the description of the present invention, the terms indicating the direction or positional relationship such as "upper", "lower", "left", "right", "inner", "outer", etc. are based on the direction or positional relationship shown in the drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention.

[0052] In addition, it should also be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0053] Please refer to Figure 1 as shown, which are respectively the step flowcharts of the interactive model display method according to the embodiments of the present invention. The method of the present invention includes:

[0054] S1, input information through the interaction unit to make the interaction unit output a display instruction and an angle instruction, and the instruction transmission unit conveys the display instruction and the angle instruction to the model construction unit;

[0055] S2, the model construction unit obtains the corresponding model information from the data storage unit based on the received display instruction and completes the construction of the model based on the model information. At the same time, based on the angle instruction, the constructed model is rotated to the corresponding angle. Among them, a single model information includes several sub-model information. The model construction unit constructs the corresponding sub-models in sequence based on each sub-model information and constrains each sub-model under the corresponding coordinates to complete the construction of the model;

[0056] S3, use the image acquisition unit to obtain the image information of the model from the corresponding angle and obtain the contour features of each sub-model at the current angle based on the image information;

[0057] S4. The analysis unit successively rates each sub-model based on the contour features of each sub-model, and after rating, determines whether the construction of the model meets the standard based on the number of sub-models at each level. When the analysis unit determines that the construction of the model does not meet the standard, it controls the correction unit to update the coordinates of the corresponding sub-model, or corrects the learning rate of model training during the model display process;

[0058] S5. When the analysis unit determines that the construction of the model meets the standard, it controls the rendering unit to start rendering the model, and after rendering is completed, outputs the rendered model to the display unit.

[0059] Please refer to Figure 2 as shown, which are respectively the module block diagrams of the interactive model display system according to the embodiments of the present invention; the system of the present invention includes:

[0060] An interaction unit for outputting a display instruction and an angle instruction based on the input information;

[0061] An instruction transmission unit connected to the interaction unit for transmitting the display instruction and the angle instruction;

[0062] A data storage unit for storing a number of model information, and each individual model information includes a number of sub-model information;

[0063] A model construction unit connected to the instruction transmission unit and the data storage unit respectively, for completing the construction of the model based on the received display instruction and the obtained corresponding model information, and rotating the constructed model to the corresponding angle based on the angle instruction;

[0064] An image acquisition unit connected to the model construction unit for acquiring the image information of the model and obtaining the contour features of each sub-model based on the image information;

[0065] An analysis unit connected to the image acquisition unit for successively rating each sub-model based on the contour features of each sub-model, and after rating, determining whether the construction of the model meets the standard based on the number of sub-models at each level. When the analysis unit determines that the construction of the model does not meet the standard, it determines the corresponding instruction;

[0066] A correction unit connected to the model construction unit and the analysis unit respectively for executing the instruction of the analysis unit, including updating the coordinates of the corresponding sub-model, or correcting the learning rate of model training during the model display process;

[0067] A rendering unit connected to the analysis unit and the model construction unit respectively for rendering the model when it is determined that the construction of the model meets the standard;

[0068] A display unit, which is connected to the rendering unit and is used to display the rendered model.

[0069] Specifically, according to the user's instruction, the model construction unit constructs the model, obtains the contour features of each sub-model, compares the contour features with the corresponding preset contour features, rates the sub-models according to the comparison results, and determines whether the construction of the model meets the standard according to the rating results. When it is determined that the construction of the model does not meet the standard, the construction parameters of the model are adjusted, including controlling the correction unit to update the coordinates of the corresponding sub-model, or correcting the learning rate of model training during the model display process, which not only effectively improves the accuracy of model generation, but also further improves the generation efficiency of the model.

[0070] In order to make the purpose and advantages of the present invention clearer and more understandable, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0071] It should be noted that the data in this embodiment are all obtained through comprehensive analysis and evaluation of the historical detection data and corresponding historical detection results of the system described in the present invention in the three months before this detection. The system described in the present invention determines the numerical values of various preset parameter standards for this detection based on 11,863 models detected cumulatively in the previous three months, the image information of each model from various angles, the determination of the corresponding angle preset contours of the sub-models in each model, the acquisition images of the actual generated contours, and the determination results. Those skilled in the art can understand that the determination method of the system described in the present invention for a single above-mentioned parameter can be to select the value with the highest proportion according to the data distribution as the preset standard parameter, use weighted summation to take the obtained value as the preset standard parameter, or other selection methods, as long as it satisfies that the system described in the present invention can clearly define different specific situations in a single determination process through the obtained values.

[0072] Specifically, based on the model construction unit, the contour features of each sub-model at the corresponding angle are determined for the preset contour features corresponding to a single sub-model, and whether the construction of the sub-model is qualified is determined based on the coincidence degree between the obtained contour features and the preset contour features, including:

[0073] If the coincidence degree is greater than the preset coincidence degree, it is determined that the construction of the sub-model is qualified, and the sub-model is recorded as a first-level sub-model;

[0074] If the coincidence degree is less than or equal to the preset coincidence degree, it is determined that the construction of the sub-model is unqualified, and the single sub-model is rated based on the dispersion amount of the coincidence abnormal area between the contour features and the preset contour features.

[0075] Specifically, it can be understood that the preset contour feature is preset to determine the specific situation of the new posture of the model trained by the model, so as to further correct the display parameters of the model, thereby improving the generation efficiency of the model.

[0076] Specifically, the preset overlap degree D0 is selected within the interval [0.75, 0.82].

[0077] Specifically, the analysis unit rates a single sub-model based on the dispersion amount of the overlapping abnormal area between the contour feature and the preset contour feature, including:

[0078] For a single sub-model, the analysis unit marks the overlapping abnormal area between the contour feature and the preset contour feature as an abnormal segment to obtain the shortest distance between each abnormal segment, and determines the ratio of the average value of each shortest distance to the average value of the lengths of each abnormal segment as the dispersion amount for this sub-model;

[0079] If the dispersion amount is less than or equal to the first preset dispersion amount, the single sub-model is recorded as a second-level sub-model;

[0080] If the dispersion amount is less than or equal to the second preset dispersion amount and greater than the first preset dispersion amount, the single sub-model is re-rated based on the number of sub-models;

[0081] If the dispersion amount is greater than the second preset dispersion amount, the single sub-model is recorded as a third-level sub-model.

[0082] Specifically, the first preset dispersion amount F1 is selected within the interval [0.25, 0.3], and the second preset dispersion amount F2 is selected within the interval [2.3, 3].

[0083] Specifically, the analysis unit re-rates a single sub-model based on the total number of sub-models, including:

[0084] If the total number is greater than the preset total number, the single sub-model is recorded as a first-level sub-model, and the current contour feature of the single sub-model is updated to the preset contour feature for the current instruction;

[0085] If the total number is less than or equal to the preset total number, the single sub-model is recorded as a third-level sub-model.

[0086] Specifically, it is selected within the interval of the preset total number N0 [3U0, 3.6U0], and U0 is the average value of the total number of sub-models corresponding to each historical display model.

[0087] Specifically, there is no limit on the selection quantity of the historical model. It can be understood that the complexity of a single model can be divided.

[0088] Specifically, when rating a single sub-model, if the overlap degree between the obtained contour features and the preset contour features is less than or equal to the preset overlap degree, there may be an abnormality in the construction of the sub-model. The comprehensive dispersion quantity is used to rate a single sub-model. The dispersion quantity represents the distribution of the overlapping abnormal areas. When the dispersion quantity is less than or equal to the first preset dispersion quantity, the overlapping abnormal areas are concentratedly distributed. In this case, there is a situation where the position of a single sub-model is shifted, resulting in partial occlusion of the sub-model. At this time, the single occluded sub-model is marked as a secondary sub-model; when the dispersion quantity is greater than the second preset dispersion quantity, the overlapping abnormal areas are dispersedly distributed. In this case, there are multiple overlapping abnormal areas in the sub-model. Since the learning rate is too large, the distance spanned by the model when updating the weights each time is too large, exceeding the range within which the model can correctly adjust to adapt to the data and the target contour, resulting in the model being continuously strengthened in the wrong area and being difficult to correct the error and move in the correct direction, thus causing abnormal phenomena of unstable and inaccurate contour fitting. At this time, the sub-model with contour features with multiple abnormalities is marked as a tertiary model; when the dispersion quantity is less than or equal to the second preset dispersion quantity and greater than the first preset dispersion quantity, the overlapping abnormal areas are evenly distributed. At this time, the total number of sub-models is determined. When the total number is greater than the preset total number, the occlusion situation in each sub-model is complex. To ensure the flexibility of model display and the tolerance for each sub-model, although the overlap degree of the contour is low, to ensure the integrity of the model architecture, its existence is accepted, and a single sub-model is determined as a primary model. The precise rating of each sub-model is carried out in turn to numerically determine the generation situation of the model, effectively improving the generation accuracy of the model.

[0089] Specifically, when the analysis unit completes the rating of each sub-model, it determines whether the construction of the model meets the standards according to the level model with the largest quantity proportion in each level of sub-models, including:

[0090] If the level model with the largest quantity proportion is a primary sub-model, it is determined that the construction of the model meets the standards, and the model is rendered;

[0091] If the level model with the largest quantity proportion is a secondary sub-model, it is determined that the construction of the model does not meet the standards, and the vertex coordinates of each secondary sub-model are corrected;

[0092] If the level model with the largest quantity proportion is a tertiary sub-model, it is determined that the construction of the model does not meet the standards, and the learning rate is adjusted to the corresponding value according to the quantity proportion of the tertiary sub-models.

[0093] Specifically, it can be understood that the quantity proportion is the ratio of the quantity of each level of sub-models to the total quantity. If there are level models with the same quantity proportion, the quantity of the model with the larger level is incremented by one. The level quantity of the tertiary sub-model is the largest.

[0094] Specifically, it is determined whether the construction of the model meets the standards according to the proportion of each level of sub-models. When the level model with the largest quantity proportion is the secondary sub-model, there is a situation where a large number of sub-models are offset in position, resulting in partial occlusion of the sub-models. The vertex coordinates of each secondary sub-model are corrected to ensure that the placement positions and hierarchical structures of the sub-models are reasonable and effectively avoid mutual occlusion. When the level model with the largest quantity proportion is the tertiary sub-model, there are instabilities and inaccuracies in the contour fitting of a large number of sub-models. At this time, the learning rate is adjusted to improve the reliability of the sub-model contour display, further improving the generation efficiency of the model.

[0095] Specifically, the correction unit is used to correct the vertex coordinates of each secondary sub-model;

[0096] For a single secondary sub-model, the correction unit is used to determine the offset according to the dispersion difference between the first preset dispersion amount and the dispersion amount, where,

[0097] The offset determined based on the dispersion difference is proportional to the dispersion difference.

[0098] In this embodiment, optionally,

[0099] The dispersion difference Y is compared with a preset first preset dispersion difference Y1 and a second preset dispersion difference Y2.

[0100] If Y≤Y1, the offset is determined as the first offset X1, and X1 = 0.1Xj is set;

[0101] If Y1<Y≤Y2, the offset is determined as the second offset X2, and X2 = 0.2Xj is set;

[0102] If Y>Y2, the offset is determined as the third offset X3, and X3 = 0.3Xj is set;

[0103] Where Xj is the initial coordinate of the j-th vertex, j = 1, 2, 3... n, n is the total number of vertices of a single sub-model, and the vertex coordinate of the corrected j-th vertex is Xj’ = Xj + Xi, i = 1, 2, 3, Y1 = 0.2F1, Y2 = 0.5F1.

[0104] Specifically, the offset is a three-dimensional vector used to represent the displacement of the sub-model. In three-dimensional space, the magnitude and direction of the offset are represented by a vector of offset components along the x-axis, y-axis, and z-axis, such as (dx, dy, dz). It can be understood that by adding the offset components on the corresponding axes to the vertex coordinates of the sub-model respectively, the correction movement of the coordinates is achieved.

[0105] Specifically, when the correction unit completes the correction of the vertex coordinates of each secondary sub-model, it re-determines whether the construction of each sub-model is qualified based on the coincidence degree between the obtained contour features of each secondary sub-model and the preset contour features. That is, if the coincidence degree is greater than the preset coincidence degree, it is determined that the construction of the sub-model is qualified, and the sub-model is recorded as a primary sub-model; if the coincidence degree is less than or equal to the preset coincidence degree, it is determined that the construction of the sub-model is unqualified, and the rotation angle of a single secondary sub-model is adjusted to the corresponding value according to the difference between the preset coincidence degree and the coincidence degree.

[0106] The correction unit is used to determine the rotation angle for a single secondary sub-model based on the coincidence difference between the preset coincidence degree and the coincidence degree. Among them,

[0107] the determined rotation angle is proportional to the coincidence difference.

[0108] In this embodiment, optionally,

[0109] the coincidence difference H is compared with the preset first preset coincidence difference H1 and the second preset coincidence difference H2.

[0110] If H ≤ H1, the rotation angle is determined as the first rotation angle R1, and it is set that R1 = 1.1R0;

[0111] If H1 < H ≤ H2, the rotation angle is determined as the second rotation angle R2, and it is set that R2 = 1.17R0;

[0112] If H > H2, the rotation angle is determined as the third rotation angle R3, and it is set that R3 = 1.23R0;

[0113] Among them, R0 is the initial rotation angle of the sub-model, H1 = 0.3D0, and H2 = 0.6D0.

[0114] Specifically, the correction unit determines the learning rate based on the proportion of the number of tertiary sub-models, where:

[0115] the learning rate determined based on the proportion of the number of tertiary sub-models is inversely proportional to the proportion of the number of tertiary sub-models.

[0116] In this embodiment, optionally,

[0117] the proportion Z of the number of tertiary sub-models is compared with the first preset proportion Z1 and the second preset proportion Z1.

[0118] If Z ≤ Z1, the learning rate is determined as the first learning rate E1, and it is set that E1 = 0.9E0;

[0119] If Z1 < Z ≤ Z2, the learning rate is determined as the second learning rate E2, and it is set that E2 = 0.82E0;

[0120] If Z > Z2, determine the learning rate as the third learning rate E3, and set E3 = 0.73E0;

[0121] where E0 is the initial learning rate, Z1 = 0.6, and Z2 = 0.8.

[0122] Specifically, during the model display process, model training is required to improve the adaptability of the data and optimize the data; during the display process, new image data will continuously be added to the display scenario, and there will be differences between these new data and the data during previous training, such as the emergence of new pose changes, etc. Through training, the model can learn and adapt to the new data features, enhancing the ability to accurately display images from different angles. It can be understood that the learning rate can control the amplitude of model weight updates. The learning rate determines the step size at which the model weights are adjusted towards a better solution during each iteration. If the learning rate is too large, the weights will oscillate violently in the wrong direction and it is difficult to converge stably to a suitable area; a too large learning rate will result in inaccurate contours. Because the too large learning rate causes the model to oscillate back and forth in the wrong area and can never effectively move in the direction of decreasing the loss function, it cannot truly find the optimal contour parameters, resulting in the contours never being able to accurately converge to the desired shape.

[0123] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the protection scope of the present invention.

[0124] The above are only the preferred embodiments of the present invention and are not used to limit the present invention; for those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An interactive model display method, characterized in that: include: Inputting information through the interaction unit so that the interaction unit outputs a display instruction and an angle instruction, and the instruction transmission unit transmits the display instruction and the angle instruction to the model building unit; The model building unit obtains corresponding model information from the data storage unit based on the received display instruction and completes the construction of the model based on the model information. At the same time, the constructed model is rotated to a corresponding angle based on the angle instruction, wherein a single model information includes a plurality of sub-model information, and the model building unit sequentially constructs corresponding sub-models based on each sub-model information, and constrains each sub-model under a corresponding coordinate to complete the construction of the model; Using an image acquisition unit to acquire image information of the model from a corresponding angle, and acquiring contour features of each sub-model at a current angle based on the image information; Using the analysis unit to sequentially rate each sub-model based on the profile features of each sub-model, and after the rating, determining whether the construction of the model meets the standard based on the number of sub-models of each grade, when the analysis unit determines that the construction of the model does not meet the standard, controlling the correction unit to update the coordinates of the corresponding sub-model, or, correcting the learning rate of the model training during the model display process; The analysis unit controls the rendering unit to start rendering the model when determining that the construction of the model meets the standard, and outputs the rendered model to the display unit after the rendering is completed; When the analysis unit completes the rating of each sub-model, it determines whether the construction of the model meets the standard according to the level model with the largest number of sub-models in each level, including: If the hierarchical model with the largest number is a first-level sub-model, it is determined that the construction of the model meets the standard and the model is rendered; If the hierarchical model with the largest number is a secondary sub-model, it is determined that the construction of the model does not meet the standard, and the coordinates of each vertex of each secondary sub-model are corrected; If the hierarchical model with the largest number is a third-level sub-model, it is determined that the construction of the model does not meet the standard, and the learning rate is adjusted to the corresponding value according to the number ratio of the third-level sub-model; The correction unit is used to correct the coordinates of each vertex of each secondary sub-model; The correction unit for a single secondary sub-model is used to determine an offset according to a dispersion difference between a first preset dispersion amount and the dispersion amount, wherein: The offset determined based on the dispersion difference is proportional to the dispersion difference.

2. The interactive model display method according to claim 1, characterized in that: The analysis unit determines the contour features of each sub-model at a corresponding angle based on the model construction unit, and determines whether the construction of the sub-model is qualified based on the overlap between the acquired contour features and the preset contour features corresponding to the single sub-model, including: If the overlap degree is greater than the preset overlap degree, the sub-model is judged to be qualified, and the sub-model is recorded as a first-level sub-model; If the degree of overlap is less than or equal to the preset degree of overlap, the construction of the sub-model is determined to be unqualified, and the individual sub-models are rated based on the dispersion of the abnormal area of ​​overlap between the contour feature and the preset contour feature.

3. The interactive model display method according to claim 2, characterized in that: The analysis unit rates a single sub-model based on the dispersion of the abnormal area where the contour feature overlaps with the preset contour feature, including: For a single sub-model, the analysis unit marks an abnormal region where a contour feature overlaps with a preset contour feature as an abnormal segment, so as to obtain the shortest distance between each abnormal segment, and determines the ratio of the average value of each shortest distance to the average value of the length of each abnormal segment as the dispersion amount for the sub-model; If the dispersion amount is less than or equal to the first preset dispersion amount, the single sub-model is recorded as a secondary sub-model; If the dispersion is less than or equal to the second preset dispersion and greater than the first preset dispersion, re-ranking the single sub-model based on the total number of sub-models; If the dispersion amount is greater than the second preset dispersion amount, the single sub-model is recorded as a third-level sub-model.

4. The interactive model display method according to claim 3, characterized in that: Individual sub-models were re-rated based on the total number of sub-models, including: If the total number is greater than the preset total number, the single sub-model is recorded as a first-level sub-model, and the current contour feature of the single sub-model is updated to the preset contour feature for the current instruction; If the total number is less than or equal to the preset total number, the single sub-model is recorded as a third-level sub-model.

5. The interactive model display method according to claim 4, characterized in that: When the correction unit completes the correction of the vertex coordinates of each secondary sub-model, it determines whether the construction of each sub-model is qualified based on the overlap between the acquired contour features of each secondary sub-model and the preset contour features, including: if the overlap is greater than the preset overlap, then the construction of the sub-model is determined to be qualified, and the sub-model is recorded as a primary sub-model; If the overlap is less than or equal to the preset overlap, the construction of the sub-model is judged to be unqualified, and the rotation angle of the single secondary sub-model is adjusted to the corresponding value according to the difference between the preset overlap and the overlap; The correction unit is used to determine the rotation angle for a single secondary sub-model based on the preset overlap degree and the overlap difference of the overlap degree, wherein: The determined rotation angle is proportional to the difference in coincidence.

6. The interactive model display method according to claim 5, characterized in that: The correction unit determines the learning rate based on the proportion of the number of three-level sub-models, wherein: The learning rate determined based on the ratio of the number of three-level sub-models is inversely proportional to the ratio of the number of three-level sub-models.

7. An interactive model display system using the method according to any one of claims 1 to 6, characterized in that: include: An interactive unit, which is used to output a display instruction and an angle instruction based on input information; An instruction transmission unit, connected to the interaction unit, for transmitting display instructions and angle instructions; A data storage unit, which is used to store a plurality of model information, wherein a single model information includes a plurality of sub-model information; a model building unit, which is connected to the instruction transmission unit and the data storage unit respectively; An image acquisition unit, connected to the model building unit, for acquiring image information of the model and acquiring contour features of each sub-model based on the image information; an analysis unit connected to the image acquisition unit; A correction unit, which is connected to the model building unit and the analysis unit respectively, and is used to execute the instructions of the analysis unit, including updating the coordinates of the corresponding sub-model, or correcting the learning rate of the model training during the model display process; A rendering unit, connected to the analysis unit and the model building unit, respectively, for rendering the model when it is determined that the building of the model meets the standard; A display unit is connected to the rendering unit and is used to display the rendered model.

Images