A motion modeling and recognition method and system for scientific visualization

Abstract

This invention discloses a method and system for motion modeling and recognition for scientific visualization, comprising: acquiring scientific time-varying data; extracting features from the scientific time-varying data; quantifying and analyzing the extracted features to obtain a set of attributes for the features; matching adjacent time slices to obtain spatiotemporal relationships; and detecting the behavior of individual features during motion. Based on the hierarchical description of the motion process and the behavior of the scientific time-varying data features, a hierarchical model is constructed; the calculation results of the scientific time-varying data features are input into the hierarchical model; the state transition probability of each feature and the settings of the hierarchical model are calculated to determine the state transition of the features and identify motion instances that conform to the definition. A motion instance database is constructed based on the motion instances, and visualization methods are used to analyze the motion in the scientific time-varying data from different levels and perspectives. This invention emphasizes the study and visualization of dynamic characteristics in scientific data, which is helpful for the observation and analysis of natural phenomena.

Description

Technical Field

[0001] This invention relates to the field of data visualization technology, and in particular to a motion modeling and recognition method and system for scientific visualization. Background Technology

[0002] In recent years, scientific visualization has been widely applied to the dynamics of various physical and chemical processes, such as combustion, ocean currents, hurricanes, and earthquakes. It effectively reveals the characteristics and laws governing these processes, playing a crucial role in assisting scientists in analyzing natural phenomena and significantly contributing to the advancement of natural sciences. Most phenomena in nature, such as combustion, ocean currents, hurricanes, and earthquakes, are essentially physical or chemical processes that continuously evolve and change in space over time. The dynamics of these processes greatly drives the progress of natural sciences. These physical and chemical processes can be expressed as differential equations containing time variables. Solving such differential equations generates data fields containing time variables, which are called three-dimensional time-varying volume data. With the rapid development of scientific computing, computer simulation, and modern observation technologies, high-resolution time-varying volume data is being generated at an increasingly rapid pace, enabling the simulation or observation of natural phenomena with unprecedented precision. These scientific datasets typically reach trillions of bytes, containing thousands of multivariable timesteps, and each timestep's data frame often has millions or even hundreds of millions of voxels. Effectively visualizing these massive datasets and further extracting key information from them has become an increasingly important challenge for many scientific research efforts.

[0003] Existing methods for visualizing scientific data can be broadly categorized into two types: time-slice visualization methods and feature-based visualization methods. Time-slice visualization methods are generally more versatile, often independent of data and expert knowledge, and are therefore supported by most commercial visualization software such as ParaView, Visit, and EnSight. However, these methods typically fail to highlight information crucial to scientific research and do not consider the time-varying and dynamic characteristics of the data, making it difficult to analyze changes within the data. Feature-based visualization methods, on the other hand, can reduce the amount of data in the visualization without losing important information. They allow users to ignore most redundant, unimportant, or uninteresting content, focusing their attention on the movement of features within the dataset, thus becoming an effective technique for studying spatiotemporal changes. However, when a motion process lasts for tens or even hundreds of time slices and overlaps spatiotemporally, observing the motion of a single feature or identifying a particular type of motion from thousands of features in the data remains extremely difficult. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides a motion modeling and recognition method and system for scientific visualization to solve the problems that existing graphical modeling cannot handle the uncertainty of short-term events in motion state transitions, and is limited by semantic limitations, making it difficult to provide accurate motion recognition strategies and motion visualization.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] In a first aspect, embodiments of the present invention provide a motion modeling and recognition method for scientific visualization, comprising: acquiring scientific time variant data, performing data feature calculation on the scientific time variant data, extracting scientific time variant data features, quantifying and analyzing the extracted features to obtain a set of attributes of the features, matching adjacent time slices to obtain spatiotemporal relationships, and detecting the behavior of individual features during motion.

[0009] A hierarchical model is constructed based on the hierarchical description of the motion process and the behavior of scientific time-varying data characteristics. The upper layer of the hierarchical model uses Petri nets to represent the motion process and obtains the transition conditions and probabilities of the motion state through the lower layer Bayesian network.

[0010] The scientific time variant data feature calculation results are input into the hierarchical model to calculate the state transition probability of each feature and the setting of the hierarchical model, determine the state transition of the feature and identify motion instances that meet the definition;

[0011] A motion instance database is constructed based on the motion instances, and motion in scientific time-varying data is analyzed from different levels and perspectives using visualization methods.

[0012] As a preferred embodiment of the motion modeling and recognition method for scientific visualization described in this invention, the data feature calculation includes: data preprocessing, feature extraction, feature tracking, and meta-event detection;

[0013] The data preprocessing includes at least data cleaning, data format transformation, variable extraction and calculation;

[0014] The feature extraction is based on the definition of the region of interest by the data expert. All voxels in a time slice are divided into foreground and / or background classes. The region growing method is used to extract the features of the coherent region composed of foreground voxels, and a set of physical and geometric properties of the features are calculated to complete the feature extraction of a time slice.

[0015] The feature extraction process iterates through each time slice in the data to extract features.

[0016] As a preferred embodiment of the motion modeling and recognition method for scientific visualization described in this invention, it further includes:

[0017] The feature tracking obtains feature point coordinate metadata based on the feature extraction. Features in adjacent time slices are compared sequentially, and the correlation between each feature is determined using a volume overlap-based method.

[0018] For any two features A and B that have volume overlap at adjacent times i and i+1, their matching degree R is expressed as:

[0019]

[0020] Among them, Vol(O i A ) represents the volume of feature A at time i, Vol(O i+1 B () represents the volume of feature B at time i+1;

[0021] Calculate the R-values ​​for different feature combinations in turn, and find the feature pairs that maximize the R-value;

[0022] Based on the mapping relationship of feature matching, the evolution process of any feature is divided into 5 states: birth, continuation, merging, splitting, and disappearance;

[0023] The results of feature tracking are recorded in a feature association table in the form of feature evolution history.

[0024] As a preferred embodiment of the motion modeling and recognition method for scientific visualization described in this invention, the hierarchical model includes: a short-time event layer model and a motion layer model;

[0025] The short-time event layer model includes a description of each short-time event, with a root node as the output of the Bayesian network;

[0026] The meta-events that make up the event description are added as input nodes to the Bayesian network;

[0027] Each input node generates an output probability to indicate its contribution to the output.

[0028] Incorporate the spatial relationships between meta-events into Bayesian networks;

[0029] Add the logical connections between meta-events to the Bayesian network;

[0030] Calculate the output of the Bayesian network, which is the probability of short-term events occurring.

[0031] As a preferred embodiment of the motion modeling and recognition method for scientific visualization described in this invention, the motion layer model includes:

[0032] Add a start state and a final state to the Petri net;

[0033] Determine the various states of motion from the motion description and add them to a Petri net;

[0034] The short-term events of state transition are determined from the motion description, and the corresponding Bayesian network short-term event model is added to the Petri net as a conditional state transition, with a preset transition condition threshold.

[0035] Add state transitions that reflect temporal relationships to Petri nets;

[0036] Add state transitions that reflect other connections to the Petri net;

[0037] Connect the various motion states and state transitions in the Petri net.

[0038] As a preferred embodiment of the motion modeling and recognition method for scientific visualization described in this invention, the method includes: calculating the state transition probability of each feature and setting the hierarchical model to determine the state transition of the feature and identify motion instances that conform to the definition, including:

[0039] The constructed motion layer model is applied to the metadata and meta-events generated by feature visualization, wherein the metadata includes features, attributes, and associated information;

[0040] Features are represented in the model as tokens;

[0041] The motion recognition algorithm iterates through all time slices in sequence. For each time slice, it first scans the behavior of all features. When a new generation, split, or merge occurs, the corresponding number of tokens is increased or decreased.

[0042] If a feature is connected to a conditional state transition, then whether the token can reach the next state is determined by calculating whether the output of the Bayesian network is greater than the state transition threshold.

[0043] When any token reaches the final state, it indicates that a movement has occurred, and a movement instance is obtained at this time.

[0044] As a preferred embodiment of the motion modeling and recognition method for scientific visualization described in this invention, the visualization method includes at least temporal visualization, spatial visualization, state induction visualization, data induction visualization, and variable visualization.

[0045] Secondly, the present invention provides a motion modeling and recognition system for scientific visualization, comprising:

[0046] The data feature calculation module is used to acquire scientific time variant data, perform data feature calculation on the scientific time variant data, extract scientific time variant data features, quantify and analyze the extracted features to obtain a set of features, match adjacent time slices to obtain spatiotemporal relationships, and detect the behavior of individual features during motion.

[0047] The model building module is used to construct a hierarchical model based on the hierarchical description of the motion process and the behavior of scientific time-varying data features. The upper layer of the hierarchical model uses Petri nets to represent the motion process and obtains the transition conditions and probabilities of the motion state through the lower layer Bayesian network.

[0048] The motion representation recognition module is used to input the scientific time variant data feature calculation results into the hierarchical model, calculate the state transition probability of each feature and the setting of the hierarchical model, determine the state transition of the feature and identify motion instances that meet the definition.

[0049] The motion visualization module is used to build a motion instance database based on the motion instances, and to analyze the motion in scientific time-varying data from different levels and perspectives using visualization methods.

[0050] Thirdly, the present invention provides a computing device, comprising:

[0051] Memory and processor;

[0052] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the motion modeling and recognition method for scientific visualization.

[0053] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the motion modeling and recognition method for scientific visualization.

[0054] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention places feature extraction, feature tracking, motion detection, and visual interaction technologies within the same methodological framework, providing a one-stop solution for observing and exploring scientific data from the time domain, spatial domain, and parameter space; by studying motion in scientific data and highlighting the dynamic characteristics of the data in visualization, it can provide accurate motion recognition strategies to help scientists better understand the occurrence and evolution of events of interest in the data; it can analyze motion in data from different levels and angles, while keeping the amount of data in visual mapping at a low level to prevent visual congestion; it highlights the dynamic characteristics of scientific data research and visualization, which is helpful for the observation and analysis of natural phenomena; and it provides a new method and tool for scientific visualization, requiring no client installation and lowering the barrier to entry. Attached Figure Description

[0055] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0056] Figure 1 This is a flowchart illustrating a motion modeling and recognition method and system for scientific visualization according to an embodiment of the present invention.

[0057] Figure 2 This is a visualization of ocean eddy motion results from ocean simulation data, as described in one embodiment of the motion modeling and recognition method and system for scientific visualization according to an embodiment of the present invention.

[0058] Figure 3 This is a visualization result of the wall turbulence simulation data of a motion modeling and recognition method and system for scientific visualization, as described in one embodiment of the present invention, on the motion of a package structure. Figure 3 (a) in the diagram is the state diagram of the motion model. Figure 3 (b) in the diagram is a schematic diagram of the movement of the package structure. Figure 3 (c) in the data represents a time slice. Detailed Implementation

[0059] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0060] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0061] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0062] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0063] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0064] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0065] Example 1

[0066] Reference Figure 1 This is one embodiment of the present invention, which provides a motion modeling and recognition method for scientific visualization, including:

[0067] S1: Acquire scientific time variant data, perform data feature calculation on the scientific time variant data, extract scientific time variant data features, quantify and analyze the extracted features to obtain a set of features, match adjacent time slices to obtain spatiotemporal relationships, and detect the behavior of individual features during the motion process;

[0068] Furthermore, data feature computation includes: data preprocessing, feature extraction, feature tracking, and meta-event detection;

[0069] Data preprocessing includes at least data cleaning, data format transformation, variable extraction and calculation;

[0070] Feature extraction is based on the definition of the region of interest by data experts. All voxels in a time slice are divided into foreground and / or background classes. The region growing method is used to extract the features of the coherent region composed of foreground voxels. A set of physical and geometric properties of the features are calculated to complete the feature extraction of a time slice.

[0071] The feature extraction process iterates through each time slice in the data to extract features.

[0072] Furthermore, this also includes:

[0073] Feature tracking is based on feature point coordinate metadata obtained from feature extraction. Features in adjacent time slices are compared sequentially, and the correlation between features is determined using a volume overlap-based method.

[0074] For any two features A and B that have volume overlap at adjacent times i and i+1, their matching degree R is expressed as:

[0075]

[0076] Among them, Vol(O i A ) represents the volume of feature A at time i, Vol(O i+1 B () represents the volume of feature B at time i+1;

[0077] Calculate the R-values ​​for different feature combinations in turn, and find the feature pairs that maximize the R-value;

[0078] Based on the mapping relationship of feature matching, the evolution process of any feature is divided into 5 states: birth, continuation, merging, splitting, and disappearance;

[0079] The results of feature tracking are recorded in a feature association table in the form of feature evolution history.

[0080] S2: Based on the hierarchical description of the motion process and the behavior of scientific time-varying data features, a hierarchical model is constructed. The upper layer of the hierarchical model uses Petri nets to represent the motion process and obtains the transition conditions and probabilities of the motion state through the lower layer Bayesian network.

[0081] Furthermore, the hierarchical model includes: a short-term event layer model and a motion layer model;

[0082] The short-time event layer model includes a description of each short-time event, with a root node as the output of the Bayesian network;

[0083] The meta-events that make up the event description are added as input nodes to the Bayesian network;

[0084] Each input node generates an output probability to indicate its contribution to the output.

[0085] Incorporate the spatial relationships between meta-events into Bayesian networks;

[0086] Add the logical connections between meta-events to the Bayesian network;

[0087] Calculate the output of the Bayesian network, which is the probability of short-term events occurring.

[0088] Specifically, the probability of a short-term event occurring is expressed as:

[0089]

[0090] Among them, P(A) i / H1) is the output probability, H1 is the root node, A 1,2,…,N This is the input node.

[0091] In an optional embodiment,

[0092] Furthermore, the motion layer model includes:

[0093] Add a start state and a final state to the Petri net;

[0094] Determine the various states of motion from the motion description and add them to a Petri net;

[0095] The short-term events of state transition are determined from the motion description, and the corresponding Bayesian network short-term event model is added to the Petri net as a conditional state transition, with a preset transition condition threshold.

[0096] Add state transitions that reflect temporal relationships to Petri nets;

[0097] Add state transitions that reflect other connections to the Petri net;

[0098] Connect the various motion states and state transitions in the Petri net.

[0099] S3: Input the scientific time variant data feature calculation results into the hierarchical model, calculate the state transition probability of each feature and the settings of the hierarchical model, determine the state transition of the feature and identify motion instances that meet the definition;

[0100] Furthermore, the state transition probability of each feature and the hierarchical model settings are calculated to determine the state transitions of the features and identify motion instances that conform to the definition, including:

[0101] The constructed motion layer model is applied to the metadata and meta-events generated by feature visualization. The metadata includes features, attributes, and related information.

[0102] Features are represented in the model as tokens;

[0103] The motion recognition algorithm iterates through all time slices in sequence. For each time slice, it first scans the behavior of all features. When a new generation, split, or merge occurs, the corresponding number of tokens is increased or decreased.

[0104] If a feature is connected to a conditional state transition, then whether the token can reach the next state is determined by calculating whether the output of the Bayesian network is greater than the state transition threshold.

[0105] When any token reaches the final state, it indicates that a movement has occurred, and a movement instance is obtained at this time.

[0106] Specifically, the state of the motion layer model is represented using Petri net states. A Petri net can be represented as:

[0107] PN=(P,T,I,O,C P ,C T ,E,M).

[0108] Where P = {P1, P2, P3, ..., P} n} represents a series of states during the motion process, C P ={C P1 ,C P2 ,C P3 ,…,C Pn} represents the condition of the state, T = {T1, T2, T3, ..., T} m} represents a series of changes in motion states, C T ={C T1 ,C T2 ,C T3 ,…,C Tm} represents the conditions for the activation of state transitions, I and O are the connecting arcs for entering and leaving the state, respectively, and M = {μ1, μ2, μ3, ..., μ n} represents the system identifier at the current moment, and is an n-dimensional variable that reflects the distribution of tokens in different states. E represents the number of times each transition is triggered.

[0109] The state update of the motion layer model is a time function, which can be written as a tuple: (M k-1 +,F);

[0110] Among them, M k-1 + It is the identifier obtained from the previous time step, and F is the update function.

[0111] The update function F updates the identifier of existing tokens in the Petri net. The update function is expressed as:

[0112]

[0113] in, The motion layer model at time t k The initial identifier.

[0114] Once all states have been updated, analyze whether each token meets the transition triggering conditions;

[0115] The excitation condition is expressed as:

[0116]

[0117] Among them, E k Indicates at t k The number of times each transition is triggered. The motion layer model at time t k The end marker.

[0118] S4: Construct a motion instance database based on motion instances, and combine visualization methods to analyze motion in scientific time variant data from different levels and perspectives;

[0119] Furthermore, the key information of each motion instance is stored in the form of a data structure, including at least the time slice and feature number, the short-term events contained therein, and the motion state, and a series of operations are provided to support querying, accessing and editing motion instances.

[0120] Based on user actions, visualization methods should be provided, including at least time-series visualization, spatial visualization, state induction visualization, data induction visualization, and variable visualization.

[0121] The above is an illustrative scheme of a motion modeling and recognition method for scientific visualization according to this embodiment. It should be noted that the technical solution of this motion modeling and recognition system for scientific visualization and the technical solution of the aforementioned motion modeling and recognition method for scientific visualization belong to the same concept. Details not described in detail in the technical solution of the motion modeling and recognition system for scientific visualization in this embodiment can be found in the description of the technical solution of the aforementioned motion modeling and recognition method for scientific visualization.

[0122] This embodiment provides a motion modeling and recognition system for scientific visualization, comprising:

[0123] The data feature calculation module is used to acquire scientific time variant data, perform data feature calculation on the scientific time variant data, extract scientific time variant data features, quantify and analyze the extracted features to obtain a set of features, match adjacent time slices to obtain spatiotemporal relationships, and detect the behavior of individual features during motion.

[0124] The model building module is used to construct a hierarchical model based on the hierarchical description of the motion process and the behavior of scientific time-varying data features. The upper layer of the hierarchical model uses Petri nets to represent the motion process and obtains the transition conditions and probabilities of the motion state through the lower layer Bayesian network.

[0125] The motion representation recognition module is used to input the scientific time variant data feature calculation results into the hierarchical model, calculate the state transition probability of each feature and the settings of the hierarchical model, determine the state transition of the feature and identify motion instances that meet the definition.

[0126] The motion visualization module is used to build a motion instance database based on motion instances and to analyze motion in scientific time-varying data from different levels and perspectives using visualization methods.

[0127] This embodiment also provides a computing device suitable for motion modeling and recognition methods for scientific visualization, including:

[0128] The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to implement the motion modeling and recognition method for scientific visualization as proposed in the above embodiments.

[0129] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements the motion modeling and recognition method for scientific visualization as proposed in the above embodiments.

[0130] The storage medium proposed in this embodiment and the motion modeling and recognition method for scientific visualization proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0131] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0132] Example 2

[0133] Reference Figures 2-3 This is one embodiment of the present invention. Unlike the first embodiment, this embodiment provides a specific application test of a motion modeling and recognition method for scientific visualization, verifying the beneficial effects of our invention.

[0134] Through the representation, identification, and visualization of motion in ocean simulation data and wall turbulence simulation data, such as Figure 3 As shown, S1 represents the first state of the motion process, S2 represents the second state, S3 represents the third state, and S4 represents the fourth state. Combined with... Figure 2 and Figure 3 As can be seen, the embodiments of the present invention extend the research on video motion recognition commonly found in the field of computer vision to the field of scientific visualization. Combining the specific implementation of visualization of ocean simulation data and wall turbulence simulation time-varying data, a hierarchical motion model is proposed to assist scientists in representing the motion process to be observed in a graphical way. Furthermore, the occurrence of ocean and wall turbulence motion is identified from ocean simulation data and wall turbulence simulation time-varying data and visualized.

[0135] This invention, based on a hierarchical motion description, uses Petri nets to represent the motion process at the upper layer of the model. It then uses a lower-level short-time event Bayesian network combined with feature calculations from time-varying data to obtain the transition conditions and probabilities of the motion state. This approach balances the uncertainty handling capabilities of Bayesian networks with the intuitive and concise process representation of Petri nets. The motion process of the features is treated as an entity, and the detected motion instances in the data are stored in a data structure for easy access, querying, and editing. Four visualization methods are provided to help scientists observe and summarize the dynamic characteristics of time-varying data from four perspectives: time, space, motion state, and statistics. This highlights the study and visualization of dynamic characteristics in scientific data, contributing to the observation and analysis of natural phenomena.

[0136] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A motion modeling and recognition method for scientific visualization, characterized in that, include: Acquire scientific time variant data, perform data feature calculations on the scientific time variant data, extract scientific time variant data features, quantify and analyze the extracted features to obtain a set of features, match adjacent time slices to obtain spatiotemporal relationships, and detect the behavior of individual features during motion. A hierarchical model is constructed based on the hierarchical description of the motion process and the behavior of scientific time-varying data characteristics. The upper layer of the hierarchical model uses Petri nets to represent the motion process and obtains the transition conditions and probabilities of the motion state through the lower layer Bayesian network. The hierarchical model includes: a short-time event layer model and a motion layer model. The short-time event layer model includes a description of each short-time event, with a root node as the output of the Bayesian network; The meta-events that make up the event description are added as input nodes to the Bayesian network; Each input node generates an output probability to indicate its contribution to the output. Incorporate the spatial relationships between meta-events into Bayesian networks; Add the logical connections between meta-events to the Bayesian network; Calculate the output of the Bayesian network, which is the probability of short-term events occurring; The motion layer model includes: Add a start state and a final state to the Petri net; Determine the various states of motion from the motion description and add them to a Petri net; The short-term events of state transition are determined from the motion description, and the corresponding Bayesian network short-term event model is added to the Petri net as a conditional state transition, with a preset transition condition threshold. Add state transitions that reflect temporal relationships to Petri nets; Add state transitions that reflect other connections to the Petri net; Connect the various motion states and state transitions in the Petri net; input the scientific time-varying data feature calculation results into the hierarchical model, calculate the state transition probability of each feature and the settings of the hierarchical model, determine the state transition of the feature and identify motion instances that meet the definition, including: The constructed motion layer model is applied to the metadata and meta-events generated by feature visualization, wherein the metadata includes features, attributes, and associated information; Features are represented in the model as tokens; The motion recognition algorithm iterates through all time slices in sequence. For each time slice, it first scans the behavior of all features. When a new generation, split, or merge occurs, the corresponding number of tokens is increased or decreased. If a feature is connected to a conditional state transition, then whether the token can reach the next state is determined by calculating whether the output of the Bayesian network is greater than the state transition threshold. When any token reaches the final state, it indicates that a movement has occurred, and a movement instance is obtained at this time; A motion instance database is constructed based on the motion instances, and motion in scientific time-varying data is analyzed from different levels and perspectives using visualization methods.

2. The motion modeling and recognition method for scientific visualization as described in claim 1, characterized in that, The data feature calculation includes: data preprocessing, feature extraction, feature tracking, and meta-event detection; The data preprocessing includes at least data cleaning, data format transformation, variable extraction and calculation; The feature extraction is based on the definition of the region of interest by the data expert. All voxels in a time slice are divided into foreground and / or background classes. The region growing method is used to extract the features of the coherent region composed of foreground voxels, and a set of physical and geometric properties of the features are calculated to complete the feature extraction of a time slice. The feature extraction process iterates through each time slice in the data to extract features.

3. The motion modeling and recognition method for scientific visualization as described in claim 2, characterized in that, Also includes: The feature tracking obtains feature point coordinate metadata based on the feature extraction. Features in adjacent time slices are compared sequentially, and the correlation between each feature is determined using a volume overlap-based method. For any two features A and B that have volume overlap at adjacent times i and i+1, their matching degree R is expressed as: Among them, Vol(O i A ) represents the volume of feature A at time i, Vol(O i+1 B () represents the volume of feature B at time i+1; Calculate the R-values ​​for different feature combinations in turn, and find the feature pairs that maximize the R-value; Based on the mapping relationship of feature matching, the evolution process of any feature is divided into 5 states: birth, continuation, merging, splitting, and disappearance; The results of feature tracking are recorded in a feature association table in the form of feature evolution history.

4. The motion modeling and recognition method for scientific visualization as described in claim 3, characterized in that: Visualization methods include at least time-series visualization, spatial visualization, state-induction visualization, data-induction visualization, and variable visualization.

5. A motion modeling and recognition system for scientific visualization, applied to the method described in any one of claims 1-4, characterized in that, include: The data feature calculation module is used to acquire scientific time variant data, perform data feature calculation on the scientific time variant data, extract scientific time variant data features, quantify and analyze the extracted features to obtain a set of features, match adjacent time slices to obtain spatiotemporal relationships, and detect the behavior of individual features during motion. The model building module is used to construct a hierarchical model based on the hierarchical description of the motion process and the behavior of scientific time-varying data features. The upper layer of the hierarchical model uses Petri nets to represent the motion process and obtains the transition conditions and probabilities of the motion state through the lower layer Bayesian network. The motion representation recognition module is used to input the scientific time variant data feature calculation results into the hierarchical model, calculate the state transition probability of each feature and the setting of the hierarchical model, determine the state transition of the feature and identify motion instances that meet the definition. The motion visualization module is used to build a motion instance database based on the motion instances, and to analyze the motion in scientific time-varying data from different levels and perspectives using visualization methods.

6. An electronic device, comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the motion modeling and recognition method for scientific visualization as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the motion modeling and recognition method for scientific visualization as described in any one of claims 1 to 4.

Images