A method for selecting virtual scene object groups based on probability propagation and position rearrangement

Abstract

The embodiments of the present disclosure disclose a method for selecting a group of virtual scene objects based on probability propagation and position rearrangement. A specific implementation of the method includes: constructing a spatial grid to accelerate the probability propagation process, querying the minimum side length of the collision body among all objects, and setting it as the side length of the smallest unit of the spatial grid; presetting the grouping center position to ensure that the position between each other is the maximum value; recording a grouping probability vector for each object; detecting the user's selection operation for ungrouped objects; based on the probability propagation algorithm of two strategies, updating the grouping probability vector and determining the grouping priority of the remaining objects to be selected; based on the object position rearrangement algorithm of the grouping probability vector, relocating the position of the target object. This implementation can increase the speed at which users classify randomly arranged multiple objects in a virtual reality scene, reduce the number of operations and difficulty, and improve the user experience.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of virtual reality interaction technology, and specifically to a method for selecting virtual scene object groups based on probability propagation and position rearrangement. Background Technology

[0002] The interaction between users and virtual objects is central to creating immersive interactive experiences in virtual reality environments. Group selection, a hot topic in this field, is a key component. Group selection is an important method for object selection, allowing users to quickly group multiple objects and significantly improving the efficiency of handling various object types.

[0003] Currently, virtual reality object selection solutions are mainly divided into single-object selection and multi-object selection. Single-object selection includes ray-based methods, as well as multi-object selection methods such as multi-ray selection, point cloud selection, and lasso selection. Group selection was first proposed in research on 2D screen virtual environments used for 3D modeling applications. Unlike the extensive exploration of the first two object selection methods, group selection schemes in virtual environments have been rarely studied previously. Some researchers have performed post-alignment tasks on grouped objects.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the inventive concept and therefore it may contain information that does not form the prior art that is already known in this country to a person of ordinary skill in the art. Summary of the Invention

[0005] The content of this disclosure is used to briefly introduce concepts that will be described in detail in the detailed description section below. The content of this disclosure is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] Some embodiments of this disclosure propose a virtual scene object group selection method based on probability propagation and position rearrangement to solve one or more of the technical problems mentioned in the background section above.

[0007] Some embodiments of this disclosure provide a virtual scene object group selection method based on probability propagation and position rearrangement. The method includes: constructing a spatial grid to accelerate the calculation of the probability propagation process; querying the minimum side length of the collider among all objects and setting it as the side length of the smallest unit of the spatial grid; pre-setting the group center position to ensure that the positions of the objects are as far apart as possible, making it easier for users to select in subsequent rounds and reducing the occurrence of misselection; recording a group probability vector for each object; updating the group probability vector based on a probability propagation algorithm with two strategies to determine the group priority of the remaining objects to be selected; and using an object position rearrangement algorithm based on the group probability vector to reposition the target object, reducing layout randomness and facilitating the next round of selection. The two strategies mentioned above are proximity exclusion and interval propagation. The proximity exclusion strategy works as follows: after a user makes a batch selection, objects adjacent to the selected object's boundary often belong to different groups; therefore, the probability of these objects belonging to the selected object's group should be very low. The interval propagation strategy works as follows: for a grouped object, the probability that its nearest ungrouped object belongs to the same group decreases, while the probability that its next nearest ungrouped object belongs to the same group increases. Simultaneously, the influence of a grouped object on its neighboring ungrouped objects is limited by a distance threshold; the influence should decrease as the distance increases, and beyond the distance threshold, the influence drops to zero. The probability propagation algorithm based on these two strategies performs the following steps: grouping the grouped objects... The probability vector is set to a standard unit vector and will not be changed thereafter. A queue similar to that used in the breadth-first search algorithm is created and maintained. First, all grouped objects are added to the queue. For each group of objects, each neighboring grid layer is traversed outwards, and any ungrouped objects encountered are added to the queue. Starting from the first ungrouped object in the queue, its group probability vector is updated. The group probability vector of the object is amplified to ensure that the effect of probability propagation is significant. The above position reordering algorithm follows these steps: probability truncation is performed, a truncation threshold is set, and probability components below the threshold are set to 0 and do not participate in the calculation of the new position of the object. The new position of the object is calculated by weighting the group probability vector of the object and the center position of each group.

[0008] The embodiments disclosed herein have the following beneficial effects:

[0009] In the embodiments of this disclosure, object group selection is defined as a three-dimensional selection task that selects multiple objects and groups them into their respective groups. The difference between group selection and multi-object selection is that multi-object selection simply selects multiple targets from a set of candidate objects, while group selection selects all objects from a set and categorizes them. The group selection process can include single and multiple object selection processes. For complex virtual scenes where objects are scattered, both selection methods may be inefficient due to different selection environments. The embodiments of this disclosure focus on introducing new strategies and optimization algorithms to explore new group selection methods, leveraging the high efficiency of multi-object selection in complex environments to achieve fast object group selection.

[0010] The group selection method based on probability propagation and position rearrangement in the embodiments of this disclosure effectively reduces the number of user operations, thereby accelerating the group selection process, reducing user workload, and improving system availability. Regarding improving group selection efficiency, the embodiments of this disclosure mainly increase the number of batch selections in the group selection process through probability propagation and object rearrangement clustering. Each round of position rearrangement improves the overall orderliness of objects, making it easier for users to use batch selection for object group selection. Furthermore, with a larger number of groups, the embodiments of this disclosure significantly increase the time to complete the object grouping task compared to existing methods. This is because with a larger number of groups, the distribution of objects in the scene becomes more dispersed, making it more difficult to use batch selection alone. In terms of the number of operations, the embodiments of this disclosure include batch selection, which requires fewer operations than methods using single-object selection. Simultaneously, the improved object orderliness due to position rearrangement enhances the efficiency of a single batch selection. Attached Figure Description

[0011] The above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0012] Figure 1 This is a flowchart of some embodiments of the virtual scene object group selection method based on probability propagation and position rearrangement according to the present disclosure.

[0013] Figure 2 This is a schematic diagram of the rearrangement layout according to some embodiments of the virtual scene object group selection method based on probability propagation and position rearrangement of this disclosure. Detailed Implementation

[0014] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of protection of the present disclosure.

[0015] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other.

[0016] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0017] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0018] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0019] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0020] Figure 1 A flowchart 100 is shown illustrating some embodiments of a virtual scene object group selection method based on probability propagation and position rearrangement according to this disclosure. This virtual scene object group selection method based on probability propagation and position rearrangement includes the following steps:

[0021] Step 101: Construct a spatial grid to accelerate the probability propagation process. Query the minimum side length of the collider among all objects and set it as the side length l of the smallest unit of the spatial grid. min Since constructing a search graph for all objects is very time-consuming, embodiments of this disclosure choose to construct a spatial grid to index all objects in the space, while accelerating the calculation of the probability propagation process. First, the total bounding box of the n objects to be selected is calculated. Then, the minimum side length of the collider among all objects is queried and set as the side length l of the smallest unit of the spatial grid. min This ensures that each smallest unit contains at most one object.

[0022] Step 102, preset the group center positions in advance to ensure that the positions between them are the maximum value. Embodiments of the present disclosure tend to set the center on the total bounding box, as Figure 2 shown. According to the different number of groups m to which the objects to be grouped belong, four layouts for object rearrangement are designed. These four center layouts are respectively applicable to the scenarios where 0 < m ≤ 4, 4 < m ≤ 8, 8 < m ≤ 12, and 12 < m ≤ 16. When m > 16, the user is allowed to manually specify a new group center.

[0023] Step 103, record a grouping probability vector p = (p1, p2,..., p m ).) for each object. Here, m is the number of groups, and p i is the probability of predicting that the object belongs to the i-th group. Initially, set the grouping probability vectors of all objects to

[0024] Step 104, detect the user's selection operation for ungrouped objects. According to the different degrees of orderliness of the object distribution in the scene, the user can use point selection and batch selection to select objects.

[0025] Step 105, update the grouping probability vector based on the probability propagation algorithm of two strategies, and determine the grouping priority of the remaining objects to be selected. Among them, the above probability propagation algorithm based on two strategies includes proximity exclusion and interval transmission. Proximity exclusion can be: after the user makes a batch selection, the objects adjacent to the boundaries of the selected objects often belong to different groups, so the probability that these objects belong to the group to which the selected objects belong should be very low; interval transmission can be: for grouped objects, the probability that the nearest ungrouped object to it belongs to the same group as it decreases, and the probability that the second nearest ungrouped object to it belongs to the same group as it increases. At the same time, the influence of the grouped object on the adjacent ungrouped object is limited by a distance threshold, and the influence should decrease as the distance between the two increases. When the distance exceeds the distance threshold, the influence is reduced to 0.

[0026] The specific implementation steps are as follows:

[0027] The first step, initialization: set the grouping probability vector of the grouped objects to the standard unit vector e k , that is, set the component corresponding to this group in the grouping probability vector to 1, and the remaining components to 0. Here, k is the number of the group to which the object belongs, and it will not be changed hereafter.

[0028] The second step is queue maintenance: Create and maintain a queue similar to that used in the breadth-first search algorithm. First, add all grouped objects to the queue. For each group of objects, traverse every neighboring layer of the grid, adding any ungrouped objects encountered to the queue. This process continues until all ungrouped objects are in the queue. Grouped objects have the highest priority, and the closer an ungrouped object is to a grouped object in a given layer, the higher its priority in the queue.

[0029] The third step, propagation: Starting with the first ungrouped object in the queue, update its grouping probability vector using the following formula:

[0030]

[0031] Where j = i-1, P i Let be the grouping probability vector of the i-th object in the queue, ⊙ be the Hadamard product, and σ be the grouping probability vector of the i-th object in the queue. ij Let be the distance between the i-th object and the j-th object in the queue. For a vector consisting entirely of 1s, the following formula can be used to obtain it:

[0032]

[0033] Among them, (x i ,y i , z i Let be the three-dimensional position coordinates of the i-th object in the virtual environment, and λ be a user-defined parameter;

[0034] The essence of probability propagation is the distance descriptor σ. ij The design. On the one hand, as the distance between the i-th and j-th objects increases, σ ij Increase, at the same time Increase, and thus P i This increases. This reflects the proximity exclusion strategy. On the other hand, based on the above analysis, σ ij This means that the closer two objects are, the greater the difference in their grouping probabilities. This reflects the interval delivery strategy.

[0035] The fourth step, probability amplification: The group probability vectors of objects are amplified to ensure the effect of probability propagation is significant. For object i, the group probability vector... Note γ=max(α1,α2,...,α m ), has an amplification function:

[0036]

[0037] Where A() is the amplification function, and β is the amplification coefficient, which is a variable parameter, and here its value is taken as 0.8. Let γ be the group probability vector. The infinite norm, i.e. The maximum value among all elements.

[0038] Step 106: Based on the object position rearrangement algorithm of group probability vector, the position of the target object is repositioned.

[0039] The first step is probability truncation: Due to the presence of low-probability components in the grouped probability vector, directly using the original grouped probability vector to calculate the new position of the object is not very effective. Therefore, a truncation threshold is set, and probability components below this threshold are set to 0 and do not participate in the calculation of the new position of the object. For the grouped probability vector of object i... There is a truncation function:

[0040]

[0041] in, Let α be the i-th component of the new grouping probability vector of objects, and med(α) be the median of the grouping probability vector.

[0042] The second step is interpolation calculation: The new position of the object is calculated by weighting the probability vector based on the object's grouping and the center positions of each group. The new position H(x, y, z) of the object is obtained by the following formula:

[0043]

[0044] Among them, C i For the center location of each group, The i-th component of the new grouping probability vector of the objects;

[0045] If an object already exists at the determined new location, a breadth-first search algorithm is used to find the nearest empty space centered on that location as the new location for the object.

[0046] Following step 106, the following steps may also be included:

[0047] Determine if all objects have been grouped. If all objects have been grouped, the process ends. If not, repeat steps 104, 105, 106, and 107.

[0048] The embodiments of this disclosure have the following beneficial effects: The group selection method based on probability propagation and position rearrangement of the embodiments of this disclosure effectively reduces the number of user operations, thereby accelerating the group selection process, effectively reducing user workload, and improving system availability. Regarding improving group selection efficiency, the embodiments of this disclosure mainly increase the number of batch selections in the group selection process through probability propagation and object rearrangement clustering. Each round of position rearrangement improves the overall orderliness of objects, making it easier for users to use batch selection for object group selection. Furthermore, with a larger number of groups, the embodiments of this disclosure significantly increase the time to complete the object grouping task compared to existing methods. This is because with a larger number of groups, the distribution of objects in the scene becomes more dispersed, making it more difficult to use batch selection alone. In terms of the number of operations, the embodiments of this disclosure include batch selection, which requires fewer operations than methods using single object selection. Simultaneously, the improved object orderliness due to position rearrangement enhances the efficiency of a single batch selection.

[0049] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for selecting virtual scene object groups based on probability propagation and position rearrangement, comprising: Construct a spatial grid to accelerate the probability propagation process, query the minimum side length of the collider among all objects, and set it as the side length of the smallest cell in the spatial grid. ; Pre-set the group center position to ensure that the position between them is the maximum value; Record a group probability vector for each object ,in, For the number of groups, To predict which category the object belongs to The probability of grouping is initially set to the grouping probability vector of all objects. ; Detect user selection actions on ungrouped objects; The probability propagation algorithm based on two strategies updates the grouping probability vector and determines the grouping priority of the remaining objects to be selected. An object position repositioning algorithm based on grouped probability vectors is used to reposition the target object. The two strategies include: proximity exclusion and interval passing, wherein, The nearest neighbor exclusion principle is as follows: After a user makes a batch selection, objects adjacent to the boundary of the selected objects often belong to different groups. Therefore, the probability that these objects belong to the group to which the selected objects belong should be very low. Interval propagation is as follows: For a grouped object, the probability that the nearest ungrouped object belongs to the same group as it decreases, and the probability that the next nearest ungrouped object belongs to the same group as it increases. At the same time, the influence of the grouped object on the nearby ungrouped objects is limited by a distance threshold. The influence should decrease as the distance between the two increases. If the distance threshold is exceeded, the influence is reduced to 0. The probability propagation algorithm based on two strategies updates the grouping probability vector and determines the grouping priority of the remaining objects to be selected, including: Set the grouping probability vector of the grouped objects to a standard unit vector. ,in, This is the group number to which the object belongs; Create and maintain a queue used in the breadth-first search algorithm. Add all grouped objects to the queue. For each group of objects, traverse each neighboring grid layer and add the ungrouped objects encountered to the queue. Starting with the first ungrouped object in the queue, update its grouping probability vector using the following formula: ， in, , For the first in the queue Group probability vector of each object, For Hadama product, For the first in the queue The object and the first The distance between objects For a vector consisting entirely of 1s, the following formula can be used to obtain it: ， in,( , , ) is the first The three-dimensional position coordinates of an object in a virtual environment These are user-defined parameters; The group probability vector of objects is amplified, for objects Grouping probability vector ,remember It has an amplification function: ， in, It is an amplification function. This is the magnification factor, a variable parameter, which is set to 0.8 here. This is a group probability vector. for The infinite norm, i.e. The maximum value among all elements.

2. The method according to claim 1, wherein, The object position repositioning algorithm based on grouped probability vectors relocates the target object's position, including: A cutoff threshold is set, and probability components below this threshold are set to 0 and not included in the calculation of the object's new position. Grouping probability vector It has a truncation function: ， in, The first group of the object probability vector One portion, This represents the median of each component of the grouped probability vector; The new position of the object is calculated by weighting the group probability vector based on the object and the center position of each group. The following formula can be used to derive: ， in, For the center location of each group, The first group of the object probability vector One component; If an object already exists at the determined new location, a breadth-first search algorithm is used to find the nearest empty space centered on that location as the new location for the object.

Images