An information chart animation automatic generation method and system based on an animation design space

Abstract

The application discloses an information chart animation automatic generation method based on an animation design space, and comprises the following steps: obtaining user input text and a static information chart matched with the user input text; analyzing the user input text and the static information chart to obtain a comprehensive analysis result; performing template matching processing on the comprehensive analysis result by using a pre-established animation design space and an animation template library to obtain a plurality of target animation templates; performing object mapping and plot analysis processing on all the target animation templates in sequence to obtain executable animation logic; and performing rendering processing on the executable animation logic to obtain a dynamic chart animation corresponding to the static information chart. The application can solve the technical problem that it is difficult to describe the nested animation structure in the information chart due to the lack of unified definition of information chart roles such as image elements, built-in text and semantic grouping in the prior art based on manual templates or preset animation rules.

Description

Technical Field

[0001] This invention belongs to the fields of visualization, computer vision, natural language processing and automatic animation generation technology, and more specifically, relates to a method and system for automatically generating infographic animations based on animation design space. Background Technology

[0002] Infographics typically integrate visual elements, thematic icons, metaphorical illustrations, and explanatory text into traditional statistical charts, simultaneously serving the functions of data presentation and visual storytelling. In digital news, data reporting, course presentations, business reports, and new media communication scenarios, text content is often used in conjunction with static infographics to explain specific data points, emphasize a trend, or guide readers to understand specific conclusions.

[0003] To reduce the cognitive burden on readers locating target areas in high-density infographics, existing technologies typically use animation to guide readers' attention to key areas. Existing methods for automatically generating infographic animations mainly fall into the following categories: The first category is based on manual templates or preset animation rules. This method usually involves designers pre-setting animation effects such as fade-in, zoom, color changes, arrow pointing, and camera zoom, and selecting appropriate templates based on the chart type or text content. The second category is based on structured chart data. This method typically relies on SVG, vector layers, or structured data tables to generate linearly playing chart animations according to the category, position, and numerical relationships of data graphics. The third category is based on visual detection or segmentation. This method uses object detection, image segmentation, or text detection techniques to identify visual elements such as bars, lines, points, icons, and text from static charts, and then adds animation effects to the identified elements. The fourth category is based on text semantic matching. This method extracts semantic information such as trends, extreme values, and comparisons from user-input text and matches this semantic information with preset animation templates to generate animation results related to the text content.

[0004] However, the aforementioned existing methods all have some significant drawbacks: First, methods based on manual templates or preset animation rules are mainly geared towards ordinary charts or single visual objects, lacking a unified definition of the roles of infographics such as image elements, embedded text, and semantic grouping, making it difficult to describe the nested animations involving data graphics, image elements, and embedded text in infographics; Second, methods based on structured chart data typically rely on SVG, vector layers, or manually preset objects, making it difficult to automatically parse the smallest animation units that can participate in animation from complex, overlapping static infographics, resulting in insufficient automated processing capabilities for complex infographics; Third, while methods based on visual detection or segmentation can identify some chart elements, they usually remain at the level of element detection or segmentation, lacking a complete mapping mechanism from visual elements and semantic grouping to animation templates and executable animation logic, causing the animation generation process to still rely on manual design experience; Fourth, methods based on text semantic matching are insufficient in modeling the multidimensional correspondence between user input text, chart scenes, animation objects, and animation plots, making it difficult to flexibly select appropriate animation templates based on the data messages expressed by the text, resulting in insufficient matching between the generated animation and the text's intended meaning. Summary of the Invention

[0005] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a method and system for automatically generating infographic animations based on an animation design space. Its purpose is to solve the following problems: First, existing methods based on manual templates or preset animation rules lack a unified definition of infographic roles such as image elements, built-in text, and semantic grouping, making it difficult to describe the nested animation structure within infographics. Second, existing methods based on structured chart data rely on SVG, vector layers, or manually preset objects, making it difficult to automatically parse the smallest animation units from complex, overlapping static infographics. Third, existing methods based on visual detection or segmentation lack a complete mapping mechanism from visual elements and semantic grouping to animation templates and executable animation logic, resulting in animation generation still relying on manual design experience. Fourth, existing methods based on text semantic matching lack sufficient modeling of the multi-dimensional correspondence between text semantics, chart scenes, animation objects, and animation plots, leading to insufficient flexibility in animation template selection and semantic matching accuracy.

[0006] To achieve the above objectives, according to one aspect of the present invention, a method for automatically generating infographic animations based on animation design space is provided, comprising the following steps: (1) Obtain the user input text and the static infographic paired with the user input text, and parse the user input text and the static infographic to obtain a comprehensive analysis result; (2) Using the pre-established animation design space and animation template library, perform template matching processing on the comprehensive analysis results obtained in step (1) to obtain multiple target animation templates; (3) Perform object mapping and plot parsing processing on all target animation templates obtained in step (2) to obtain executable animation logic; (4) Render the executable animation logic obtained in step (3) to obtain a dynamic chart animation corresponding to the static information chart.

[0007] Preferably, step (1) specifically involves: firstly, acquiring the user input text and a static infographic paired with it; then, performing semantic analysis on the user input text to extract the message type, target data content, and target expression intent; subsequently, performing a standardization preprocessing operation on the static infographic to obtain a standardized preprocessed static infographic; then, performing fine-grained element analysis on the standardized preprocessed static infographic to obtain multiple data graphics, image elements, and embedded text; then, performing semantic grouping on the obtained data graphics, image elements, and embedded text to obtain animation object groups, wherein semantic grouping refers to grouping data graphics, image elements, and embedded text corresponding to the same data item, the same visual area, or the same expression meaning into the same animation object group; after fine-grained element analysis, the static infographic yields multiple candidate animation objects, and animation object groups are formed based on the spatial proximity and semantic association between the objects; finally, combining the obtained message type, target data content, target expression intent, and animation object groups into a comprehensive analysis result.

[0008] Preferably, step (2) specifically involves: firstly, selecting multiple candidate animation templates from a pre-established animation template library based on the message type and the chart type of the static infographic in the comprehensive analysis result; then, representing the comprehensive analysis result as an input triple I={m, v, e}, where m represents the message type corresponding to the user input text, v represents the scene type corresponding to the static infographic, and e represents the set of animation object types obtained after fine-grained element analysis and semantic grouping, the animation object type set e including one or more of data graphics, image elements, and built-in text; subsequently, for each candidate animation template t, obtaining its template triple {mt} Let mt, vt, et, be the template message type, vt be the applicable scenario type, and et be the set of animation object types required by the template. Then, calculate the message matching score Sm(t, m), scenario matching score Sv(t, v), and animation object matching score Se(t, e) of the candidate animation template t respectively, and calculate the matching score of the candidate animation template t according to Score(t) = λ1·Sm(t, m) + λ2·Sv(t, v) + λ3·Se(t, e). Finally, determine the candidate animation templates whose matching scores meet the preset conditions as the target animation templates; where λ1, λ2, and λ3 are preset weights. Wherein, when mt is the same as m, Sm(t, m) = 1, otherwise Sm(t, m) = 0; when vt is the same as v, Sv(t, v) = 1, otherwise Sv(t, v) = 0; the animation object matching score Se(t, e) is determined by the support score and the utilization score. The support score is used to represent the frequency of occurrence of the set of animation object types et required by the template in the animation template library, and the utilization score is used to represent the degree to which the set of animation object types et required by the template is covered by the set of input animation object types e; when et cannot be covered by e, Se(t, e) = 0, when et can be covered by e, Se(t, e) is equal to the sum of the support score and the utilization score; The preset conditions are: the matching score of the candidate animation template is greater than or equal to the preset matching threshold θ, the message type of the candidate animation template is consistent with the message type in the comprehensive analysis result, the applicable chart type of the candidate animation template is consistent with the chart type of the static information chart, and the applicable animation object type of the candidate animation template is covered by any object type in the animation object group; where θ ranges from 0.5 to 3. When the number of candidate animation templates that meet the above preset conditions is zero, the top K candidate animation templates with the highest matching scores are determined as the target animation templates, where K is a positive integer.

[0009] Preferably, the animation design space and animation template library in step (2) are established according to the following steps: (2-1) Obtain multiple infographic animation samples, and annotate the text content, chart type, animation object and animation process in all infographic animation samples. The annotated text content, chart type, animation object and animation process constitute the animation design sample set. (2-2) Perform multidimensional label statistics, co-occurrence count statistics and category merging on the animation design sample set obtained in step (2-1) to obtain an animation design space composed of multiple dimensions. The animation design space includes message dimension, scene dimension, animation object dimension and plot dimension. Each dimension is used to describe the data message, applicable scene, objects involved in the animation and the organization of the animation plot in the infographic animation. (2-3) Establish a tree structure set for the animation design space corresponding to the plot dimension obtained in step (2-2); the tree structure corresponding to the plot dimension is used to represent the tiling relationship or nesting relationship between animation events, where the root node represents the global rhythm type, the leaf node represents the smallest animation unit, and the internal rhythm node represents the organization method of the sub-animation units within the semantic group. (2-4) Based on the animation design space obtained in step (2-2) and the tree structure set obtained in step (2-3), an animation template library is established by using template encoding, rule merging and parameter statistics. (2-5) Generate template matching information and animation mapping information based on the animation template library obtained in step (2-4); wherein, the template matching information is used to determine the matching score of the candidate animation template from the animation template library based on the comprehensive analysis results obtained in the online animation generation process; the animation mapping information is used to map the abstract animation objects, animation effects and playback order in the target animation template to the animation objects, animation parameters and executable animation logic in the static information chart; (2-6) Using the template matching information obtained in step (2-5), the comprehensive analysis result obtained in step (1) and the animation template library obtained in step (2-4) are matched and the candidate animation templates whose matching scores meet the preset conditions are determined as the target animation templates. When the number of candidate animation templates that meet the preset conditions is zero, the top K candidate animation templates with the highest matching scores are determined as the target animation templates in descending order of matching scores, so as to obtain multiple target animation templates.

[0010] Preferably, step (2-2) specifically includes the following sub-steps: (2-2-1) Set the counter i=1 and initialize the global standard label set to empty; (2-2-2) Determine whether i is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-2-6); otherwise, proceed to step (2-2-3). (2-2-3) Obtain the message tag, scene tag, animation object tag, and animation process tag of the i-th animation design sample in the animation design sample set; Specifically, message tags are used to characterize the type of data fact conveyed by the animation design sample, scene tags are used to characterize the type of infographic corresponding to the animation design sample, animation object tags are used to characterize the data graphics, image elements, built-in text or semantic groups involved in the animation, and animation process tags are used to characterize the animation effects, rhythm, shots and sequence; among them, rhythm includes one or more of the following: individual presentation, parallel presentation, phased presentation, staggered presentation and layered presentation. (2-2-4) Normalize and merge the tags with the same meaning or the same function in the i-th animation design sample in the animation design sample set to obtain the standard tag set corresponding to the animation design sample, including message standard tag subset, scene standard tag subset, animation object standard tag subset, and animation process standard tag subset; (2-2-5) Add the set of standard labels corresponding to the i-th animation design sample obtained in step (2-2-4) to the global standard label set, set i=i+1, and return to step (2-2-2). (2-2-6) Count the frequency of occurrence of each standard label in the global standard label set, and count the number of times each pair of standard labels co-occurs in the same animation design sample; (2-2-7) Determine the candidate categories and candidate category combinations based on the frequency of occurrence and co-occurrence of the standard labels as calculated in step (2-2-6); Specifically, when the frequency of a standard tag is greater than or equal to the first frequency threshold α, the standard tag is identified as a candidate category, and the dimension to which the candidate category belongs is recorded according to the message standard tag subset, scene standard tag subset, animation object standard tag subset, or animation process standard tag subset to which the standard tag belongs; where, when the standard tag belongs to the animation process standard tag subset, its dimension is recorded as the plot dimension; the value range of the first frequency threshold α is 2 to 30. When the co-occurrence frequency of two standard tags is greater than or equal to the second frequency threshold γ, or the co-occurrence ratio of two standard tags is greater than or equal to the co-occurrence ratio threshold ρ, the two standard tags are set as a candidate category combination that satisfies the co-occurrence condition; where, the value range of the second frequency threshold γ is 2 to 20; the co-occurrence ratio is the ratio between the co-occurrence frequency of two standard tags and the smaller value of the occurrence frequency of the two standard tags, and the value range of the co-occurrence ratio threshold ρ is 0.1 to 0.8. (2-2-8) Synonym tags are merged for the candidate categories and candidate category combinations obtained in step (2-2-7), and the candidate categories and candidate category combinations are filtered according to the number of co-occurrences, the co-occurrence ratio and the frequency of category occurrence to obtain the animation design space corresponding to the message dimension, scene dimension, animation object dimension and plot dimension. Specifically, this step involves first processing each candidate category according to its dimension, without merging candidate categories from different dimensions; then, merging candidate categories with the same meaning or function within the same dimension to obtain higher-level candidate categories; subsequently, screening candidate category combinations based on their co-occurrence frequency and co-occurrence ratio. When the co-occurrence frequency of two candidate categories is greater than or equal to a second frequency threshold γ, or the co-occurrence ratio of two candidate categories is greater than or equal to a co-occurrence ratio threshold ρ, the candidate category combination is retained. The second frequency threshold γ ranges from 2 to 20; the co-occurrence ratio is the smaller of the co-occurrence frequency and the frequency of occurrence of the two candidate categories. The ratio between values, the co-occurrence ratio threshold ρ ranges from 0.1 to 0.8; when the frequency of a candidate category is less than the first frequency threshold α, and the candidate category cannot form a complete animation rule with other candidate categories, the candidate category is removed; the first frequency threshold α ranges from 2 to 30; when a candidate category combination does not meet the above co-occurrence conditions and cannot form a complete animation rule, the candidate category combination is removed; finally, the remaining candidate categories and candidate category combinations are merged according to four aspects: message, scene, animation object, and plot, to obtain the animation design space corresponding to the message dimension, scene dimension, animation object dimension, and plot dimension, respectively; Specifically, the animation design space includes message dimension, scene dimension, animation object dimension, and plot dimension; message dimension is used to represent the type of data fact to be expressed; scene dimension is used to represent the chart type of static infographics; animation object dimension is used to represent data graphics, image elements, built-in text and their semantic groupings; plot dimension is used to represent animation effects, rhythm, shots and sequence.

[0011] Preferably, step (2-3) specifically includes the following sub-steps: (2-3-1) Set the first counter j=1; (2-3-2) Determine whether j is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-3-10); otherwise, proceed to step (2-3-3). (2-3-3) Obtain the animation process label of the j-th animation design sample in the animation design sample set, and use the global rhythm type in the animation process label as the root node of the plot tree corresponding to the j-th animation design sample; (2-3-4) Obtain the set of minimum animation units in the j-th animation design sample, and sort the set of minimum animation units according to the playback start time to obtain the sorted set of minimum animation units Uj={u1, u2, ..., uMj}, where Mj represents the total number of minimum animation units in the j-th animation design sample; (2-3-5) Set the second counter s=1; (2-3-6) Determine if s is greater than Mj. If so, proceed to step (2-3-9). Otherwise, read the s-th smallest animation unit us from the sorted set of smallest animation units and create a leaf node for the smallest animation unit us. The leaf node records the animation object type, animation effect, camera effect, start time and duration. (2-3-7) Determine whether the animation object type corresponding to the s-th smallest animation unit us is a semantic grouping type. If it is, obtain the semantic grouping object corresponding to the semantic grouping type and determine whether there are more than two sub-animation units inside the semantic grouping object. If not, keep the leaf node, set s=s+1, and return to step (2-3-6). If it exists, replace the leaf node with the internal rhythm node and go to step (2-3-8). Otherwise, keep the leaf node, set s=s+1, and return to step (2-3-6). (2-3-8) Obtain the set of sub-animation units inside the semantic grouping object, and sort the set of sub-animation units according to the playback start time; under the internal rhythm node, create leaf nodes corresponding to each sub-animation unit in sequence, and record the animation object type, animation effect, camera effect, start time and duration of the corresponding sub-animation unit in each leaf node; set s=s+1, and return to step (2-3-6). (2-3-9) Combine the root node, internal rhythm node, leaf node and child leaf node that have been established in the j-th animation design sample to form the plot tree corresponding to the animation design sample; set j=j+1 and return to step (2-3-2). (2-3-10) Perform isomorphic merging on the plot trees corresponding to all animation design samples. Merge plot trees with the same root node rhythm type, leaf node animation object type, leaf node animation effect, camera method and parent-child hierarchy into the same tree structure to obtain the set of tree structures corresponding to the plot dimension.

[0012] Preferably, step (2-4) specifically includes the following sub-steps: (2-4-1) Set the counter k=1; (2-4-2) Determine whether k is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-4-7); otherwise, proceed to step (2-4-3). (2-4-3) Read the message dimension label, scene dimension label, animation object dimension label and plot tree corresponding to the kth animation design sample; (2-4-4) The template encoding of the kth animation design sample is performed in the order of "message type - applicable chart type - applicable animation object type - plot tree structure" to obtain the template encoding result of the kth animation design sample; (2-4-5) Determine whether the template encoding result of the kth animation design sample obtained in step (2-4-4) already exists in the animation template library. If so, add the animation effect parameters, rhythm parameters and camera parameters of the kth animation design sample to the parameter set of the corresponding template; otherwise, create a new animation template record and write the template encoding result into the animation template record. (2-4-6) Set k=k+1 and return to step (2-4-2); (2-4-7) Perform statistical processing on the parameter set recorded in each animation template to obtain the default parameter value range of the duration, delay time, animation direction, scaling ratio and camera operation mode of the animation template; (2-4-8) Determine whether the number of times the sample appears in the animation template record is greater than or equal to the preset number threshold η. If so, retain the animation template record; otherwise, merge the animation template record into the animation template record that is most similar to its message type, applicable chart type and plot tree structure, or mark it as a low-frequency template; where the value of η ranges from 2 to 20. (2-4-9) Assign a template number to each retained animation template record and generate an animation template library containing message type, applicable chart type, applicable animation object type, animation effect, rhythm structure, shot mode, playback order and default parameter value range; Each animation template in the animation template library includes at least the message type, applicable chart type, applicable animation object type, animation effect, rhythm structure, camera style, and playback order.

[0013] Preferably, step (3) specifically includes the following sub-steps: (3-1) Set the counter r=1; (3-2) Determine whether r is greater than the total number R of target animation templates obtained in step (2). If so, proceed to step (3-9); otherwise, proceed to step (3-3). (3-3) Obtain the plot tree structure, applicable animation object type set, animation effect set, rhythm structure, camera mode and default parameter value range in the r-th target animation template; (3-4) For each leaf node in the plot tree structure, obtain the template animation object type corresponding to the leaf node, and filter all candidate animation objects with the same object type or that can be covered by the object type from the animation object group obtained in step (1); calculate the binding score Bind(o) of the candidate animation object based on the semantic relevance between each candidate animation object and the target data content, the spatial proximity between the candidate animation object and the target expression region, and the hierarchical consistency between the group to which the candidate animation object belongs and the template group, where Bind(o) o) = μ1·Sr(o) + μ2·Sp(o) + μ3·Sh(o); Sr(o) represents the semantic relevance score between the candidate animation object o and the target data content, Sp(o) represents the spatial proximity score between the candidate animation object o and the target expression region, and Sh(o) represents the hierarchical consistency score between the group to which the candidate animation object o belongs and the template group; μ1, μ2, and μ3 are preset weights, with μ1 ranging from 0.1 to 1; μ2 ranging from 0.1 to 1; and μ3 ranging from 0.1 to 1, preferably 0.2; (3-5) For each leaf node in the plot tree structure, determine whether the number of candidate animation objects corresponding to the leaf node obtained in step (3-4) is zero; if so, mark the leaf node as an unbound node and skip the unbound node when generating executable animation logic in the future; otherwise, select the candidate animation object with the largest binding score from all candidate animation objects corresponding to the leaf node and record its binding score as the maximum binding score; determine whether the maximum binding score is greater than or equal to the preset binding threshold β; if so, bind the candidate animation object corresponding to the maximum binding score to the specific animation object corresponding to the leaf node; otherwise, mark the leaf node as an unbound node and skip the unbound node when generating executable animation logic in the future; where β ranges from 0.3 to 1; after completing the judgment of all leaf nodes in the plot tree structure, obtain the leaf node binding result corresponding to the r-th target animation template and proceed to step (3-6). (3-6) Based on the leaf node binding results obtained in step (3-5), the plot tree structure in the r-th target animation template is recursively traversed. When a bound leaf node is traversed, the animation event corresponding to that leaf node is generated. When an unbound node is traversed, the unbound node is skipped. When an internal rhythm node is traversed, the time relationship between its child nodes is determined according to the rhythm type of the internal rhythm node, thereby obtaining the animation event set and time relationship corresponding to the r-th target animation template. (3-7) Based on the set of animation events and time relationships obtained in step (3-6), assign an animation start time and duration to each animation event, and generate corresponding animation effect parameters according to the animation effect type, thereby obtaining the set of animation event parameters corresponding to the r-th target animation template, including animation event, specific animation object, animation start time, duration, animation effect parameters and camera parameters; (3-8) Convert the set of animation event parameters corresponding to the r-th target animation template obtained in step (3-7) into the executable animation sub-logic corresponding to the target animation template, set r=r+1, and return to step (3-2). (3-9) Integrate all executable animation sub-logics corresponding to the target animation templates in descending order of template matching score to obtain executable animation logic; wherein, executable animation logic includes one or more of HTML structure, CSS animation style and JavaScript timeline control code.

[0014] Preferably, in step (3-6), determining the temporal relationship between the child nodes based on the rhythm type of the internal rhythm node specifically means that when the rhythm type is presented alone, an animation event is executed independently; when the rhythm type is presented in parallel, multiple animation events have the same start time; when the rhythm type is presented in stages, multiple animation events are executed sequentially; when the rhythm type is presented in staggered manner, multiple animation events are executed sequentially at preset time intervals; and when the rhythm type is presented in layered manner, the next animation event starts before the previous animation event has completely finished. In steps (3-7), the generated effects correspond to the parameters of transparency, scaling ratio, or displacement start point; the style change effects correspond to the parameters of color, transparency, stroke, or glow; the geometric change effects correspond to the parameters of position, rotation angle, scaling ratio, or shape change; the pointing effects correspond to the generation parameters of the arrow auxiliary element; the guiding effects correspond to the generation parameters of the path line auxiliary element; the explanation effects correspond to the generation parameters of the text box auxiliary element; and the highlight effects correspond to the generation parameters of the border auxiliary element. At the same time, scaling, translation, or toggling parameters are generated according to the camera mode. Step (4) specifically involves first determining the display state of each animation object on the timeline based on the executable animation logic obtained in step (3); then, dynamically rendering the data graphics, image elements, built-in text, or combinations thereof corresponding to the animation object in the static infographic based on the display state; and finally, synthesizing the dynamic rendering results corresponding to all animation objects to obtain a dynamic infographic animation that matches the user's input text.

[0015] According to another aspect of the present invention, an automatic generation system for infographic animations based on animation design space is provided, comprising the following modules: The first module is used to obtain user input text and static infographics paired with user input text, and to parse user input text and static infographics to obtain comprehensive analysis results; The second module is used to perform template matching processing on the comprehensive analysis results obtained from the first module using a pre-established animation design space and animation template library to obtain multiple target animation templates. The third module is used to perform object mapping and plot parsing on all the target animation templates obtained from the second module in order to obtain executable animation logic; The fourth module is used to render the executable animation logic obtained from the third module to obtain a dynamic chart animation corresponding to the static infographic.

[0016] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: First, because the present invention employs steps (2-1) to (2-3), it constructs an animation design space containing message dimension, scene dimension, animation object dimension, and plot dimension by statistically analyzing, merging, and modeling the message tags, scene tags, animation object tags, and animation process tags in the infographic animation sample. Therefore, it can solve the technical problem that existing methods based on manual templates or preset animation rules lack a unified definition of infographic roles such as image elements, built-in text, and semantic grouping, making it difficult to describe the nested animation structure of infographics.

[0017] Second, because the present invention employs step (1), which parses user-input text and static infographics, and performs fine-grained element parsing and semantic grouping of data graphics, image elements and built-in text in static infographics, it can solve the technical problem that existing methods based on structured chart data are difficult to automatically parse the smallest animation unit from complex and overlapping static infographics due to their reliance on SVG, vector layers or manually preset objects.

[0018] Third, because the present invention employs steps (2-4) to (3), it converts the abstract template in the animation design space into an executable animation sub-logic containing animation events, specific animation objects, time parameters, animation effect parameters, and camera parameters through template encoding, template matching information, animation mapping information, object binding, and plot tree parsing. Therefore, it can solve the technical problem that existing visual detection or segmentation methods lack a complete mapping mechanism from visual elements and semantic grouping to animation templates and executable animation logic, resulting in animation generation still relying on manual design experience.

[0019] Fourth, this invention employs steps (2) and (3-4) to (3-5), which jointly judge the correspondence between user input text, static infographics, animation object groups, and target animation templates based on message matching scores, scene matching scores, animation object matching scores, and candidate animation object binding scores. Therefore, it can solve the technical problem of insufficient flexibility in animation template selection and semantic matching degree caused by the lack of modeling of the multidimensional correspondence between text semantics, infographic scenes, animation objects, and animation plots in existing text semantic matching methods.

[0020] Fifth, by organizing the template matching, object binding, plot parsing and animation rendering processes into a continuous automated process, the present invention allows users to obtain dynamic chart animations that match the text's intended meaning simply by inputting text and static infographics, thus reducing the labor costs and operational difficulty of creating infographic animations. Attached Figure Description

[0021] Figure 1 This is the overall flowchart of the method for automatically generating infographic animations based on animation design space according to the present invention; Figure 2 This is a schematic diagram of the animation design space in this invention; Figure 3 This is a schematic diagram of the tree structure of the plot dimension in this invention; Figure 4 This is a flowchart illustrating the fine-grained element parsing and semantic grouping of infographics in this invention. Figure 5 This is a schematic diagram of the animation template matching and executable animation mapping process in this invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0023] The basic idea of ​​this invention is as follows: First, by labeling, statistically analyzing, and merging multiple infographic animation samples, an animation design space is established, encompassing message, scene, animation object, and plot dimensions. Based on the plot dimension, a tree-structured set capable of describing tiling and nested animations is created. Second, an animation template library is generated based on the animation design space and the tree-structured set, enabling each animation template to simultaneously represent the message type, applicable chart type, applicable animation object type, and plot tree structure. Then, during online animation generation, the user-input text and static infographics are parsed to obtain a comprehensive analysis result including message type, target data content, target expression intent, and animation object grouping. Subsequently, the target animation template is determined based on the matching score between the comprehensive analysis result and the animation template library, and the abstract animation objects in the target animation template are bound to specific animation objects in the static infographic. Finally, the plot tree structure in the target animation template is parsed to generate executable animation logic and perform dynamic rendering, thereby obtaining a dynamic chart animation that matches the user-input text.

[0024] like Figure 1 As shown, this invention provides a method for automatically generating infographic animations based on animation design space, comprising the following steps: (1) Obtain the user input text and the static infographic paired with the user input text, and parse the user input text and the static infographic to obtain a comprehensive analysis result; Specifically, this step involves first acquiring the user input text and a corresponding static infographic; then, performing semantic parsing on the user input text to extract the message type, target data content, and target expressive intent; subsequently, performing standardization preprocessing on the static infographic (this standardization preprocessing includes converting the static infographic to a preset image format and resolution, unifying the image coordinate origin and scale, correcting the positional correspondence between the chart area, legend area, and text area, and removing one or more irrelevant edge areas that affect detection) to obtain a standardized preprocessed static infographic; next, performing fine-grained element parsing on the standardized preprocessed static infographic to obtain multiple data graphics, image elements, and embedded text; then, semantically grouping the obtained data graphics, image elements, and embedded text to obtain animation object groups (where semantic grouping refers to grouping data graphics, image elements, and embedded text corresponding to the same data item, the same visual area, or the same expressive meaning into the same animation object group); such as Figure 4As shown, after fine-grained element parsing, the static infographic yields multiple candidate animation objects, which are then grouped based on spatial proximity and semantic relationships. Finally, the obtained message type, target data content, target expression intent, and animation object groupings are combined into a comprehensive parsing result.

[0025] The advantage of this step (1) is that by parsing the user input text and the static information chart at the same time, it is possible to establish a correspondence between the message type, target data content, target expression intent in the text and the animation objects in the chart, providing a unified input basis for subsequent template matching and object binding.

[0026] (2) Using the pre-established animation design space and animation template library, perform template matching processing on the comprehensive analysis results obtained in step (1) to obtain multiple target animation templates; Specifically, this step involves first filtering multiple candidate animation templates from a pre-established animation template library based on the message type and the chart type of the static infographic in the comprehensive analysis results; then, representing the comprehensive analysis results as an input triple I={m, v, e}, where m represents the message type corresponding to the user input text, v represents the scene type corresponding to the static infographic, and e represents the set of animation object types obtained after fine-grained element parsing and semantic grouping, with the animation object type set e including one or more of data graphics, image elements, and built-in text; subsequently, for each candidate animation template t, its template triple is obtained. The tuple {mt, vt, et} is used, where mt represents the template message type, vt represents the applicable scenario type of the template, and et represents the set of animation object types required by the template. Then, the message matching score Sm(t, m), scenario matching score Sv(t, v), and animation object matching score Se(t, e) of the candidate animation template t are calculated respectively, and the matching score of the candidate animation template t is calculated according to Score(t) = λ1·Sm(t, m) + λ2·Sv(t, v) + λ3·Se(t, e). Finally, the candidate animation templates whose matching scores meet the preset conditions are determined as the target animation templates. Wherein, when mt is the same as m, Sm(t, m) = 1, otherwise Sm(t, m) = 0; when vt is the same as v, Sv(t, v) = 1, otherwise Sv(t, v) = 0; the animation object matching score Se(t, e) is determined by the support score and the utilization score. The support score is used to represent the frequency of occurrence of the set of animation object types et required by the template in the animation template library, and the utilization score is used to represent the degree to which the set of animation object types et required by the template is covered by the set of input animation object types e; when et cannot be covered by e, Se(t, e) = 0, when et can be covered by e, Se(t, e) is equal to the sum of the support score and the utilization score; More specifically, λ1, λ2, and λ3 are preset weights. The value of λ1 ranges from 0.1 to 1, preferably 0.5; the value of λ2 ranges from 0.1 to 1, preferably 0.5; and the value of λ3 ranges from 0.5 to 2, preferably 1. The preset conditions are: the matching score of the candidate animation template is greater than or equal to the preset matching threshold θ, and the message type of the candidate animation template is consistent with the message type in the comprehensive analysis result; the applicable chart type of the candidate animation template is consistent with the chart type of the static information chart; and the applicable animation object type of the candidate animation template is covered by any object type in the animation object group. Among these conditions, the value of θ ranges from 0.5 to 3, preferably 1.5. When the number of candidate animation templates that meet the above preset conditions is zero, the top K candidate animation templates with the highest matching scores are determined as the target animation templates in descending order of matching scores. K is a positive integer, and its value ranges from 1 to 10, preferably 3.

[0027] The animation design space and animation template library in step (2) are established according to the following steps: (2-1) Obtain multiple infographic animation samples, and annotate the text content, chart type, animation object and animation process in all infographic animation samples. The annotated text content, chart type, animation object and animation process constitute the animation design sample set. (2-2) The animation design sample set obtained in step (2-1) is subjected to multi-dimensional label statistics, co-occurrence count statistics, and category merging to obtain an animation design space composed of multiple dimensions; such as Figure 2 As shown, the animation design space includes message dimension, scene dimension, animation object dimension, and plot dimension. These dimensions work together to describe the data messages, applicable scenarios, objects involved in the animation, and the way the animation plot is organized in the infographic animation.

[0028] This step specifically includes the following sub-steps: (2-2-1) Set the counter i=1 and initialize the global standard label set to empty; (2-2-2) Determine whether i is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-2-6); otherwise, proceed to step (2-2-3). (2-2-3) Obtain the message tag, scene tag, animation object tag, and animation process tag of the i-th animation design sample in the animation design sample set; Specifically, message tags are used to characterize the type of data fact conveyed by the animation design sample, scene tags are used to characterize the type of infographic corresponding to the animation design sample, animation object tags are used to characterize the data graphics, image elements, built-in text or semantic groups involved in the animation, and animation process tags are used to characterize the animation effects, rhythm, shots and sequence; among them, rhythm includes one or more of the following: individual presentation, parallel presentation, phased presentation, staggered presentation and layered presentation. (2-2-4) Normalize and merge the tags with the same meaning or the same function in the i-th animation design sample in the animation design sample set to obtain the standard tag set corresponding to the animation design sample, including message standard tag subset, scene standard tag subset, animation object standard tag subset, and animation process standard tag subset; (2-2-5) Add the set of standard labels corresponding to the i-th animation design sample obtained in step (2-2-4) to the global standard label set, set i=i+1, and return to step (2-2-2). (2-2-6) Count the frequency of occurrence of each standard label in the global standard label set, and count the number of times each pair of standard labels co-occurs in the same animation design sample; (2-2-7) Determine the candidate categories and candidate category combinations based on the frequency of occurrence and co-occurrence of the standard labels as statistically analyzed in step (2-2-6).

[0029] Specifically, when the frequency of occurrence of a standard tag is greater than or equal to the first frequency threshold α, the standard tag is determined as a candidate category, and the dimension to which the candidate category belongs is recorded according to the message standard tag subset, scene standard tag subset, animation object standard tag subset, or animation process standard tag subset to which the standard tag belongs; wherein, when the standard tag belongs to the animation process standard tag subset, its dimension is recorded as the plot dimension; the value range of the first frequency threshold α is 2 to 30, preferably 5. When the co-occurrence frequency of two standard tags is greater than or equal to the second frequency threshold γ, or the co-occurrence ratio of two standard tags is greater than or equal to the co-occurrence ratio threshold ρ, the two standard tags are set as a candidate category combination that satisfies the co-occurrence condition; wherein, the value range of the second frequency threshold γ is 2 to 20, preferably 3; the co-occurrence ratio is the ratio between the co-occurrence frequency of two standard tags and the smaller value of the occurrence frequency of the two standard tags, and the value range of the co-occurrence ratio threshold ρ is 0.1 to 0.8, preferably 0.3; (2-2-8) Synonym tags are merged for the candidate categories and candidate category combinations obtained in step (2-2-7), and the candidate categories and candidate category combinations are filtered according to the number of co-occurrences, the co-occurrence ratio and the frequency of category occurrence to obtain the animation design space corresponding to the message dimension, scene dimension, animation object dimension and plot dimension. Specifically, this step involves first processing each candidate category according to its dimension, without merging candidate categories from different dimensions; then, merging candidate categories with the same meaning or function within the same dimension to obtain a higher-level candidate category (wherein, candidate categories with the same meaning or function refer to those that were grouped into the same standard label set in the normalization and merging process in step (2-2-4), or those used in animation design samples to represent the same type of message, the same type of chart scene, the same type of animation object, or the same type of animation process); subsequently, the candidate category combinations are screened based on their co-occurrence frequency and co-occurrence ratio. When two candidate categories co-occur... When the frequency of occurrence of a candidate category is greater than or equal to the second frequency threshold γ, or the co-occurrence ratio of two candidate categories is greater than or equal to the co-occurrence ratio threshold ρ, the candidate category combination is retained; wherein, the value of the second frequency threshold γ is in the range of 2 to 20, preferably 3; the co-occurrence ratio is the ratio between the co-occurrence frequency of two candidate categories and the smaller value of the occurrence frequency of the two candidate categories, and the value of the co-occurrence ratio threshold ρ is in the range of 0.1 to 0.8, preferably 0.3; when the occurrence frequency of a candidate category is less than the first frequency threshold α, and the candidate category cannot form a complete animation rule with other candidate categories, the candidate category is removed; wherein, the value of the first frequency threshold α is in the range of 2 to 30, preferably 5. When a candidate category combination does not meet the above co-occurrence conditions and cannot form a complete animation rule, the candidate category combination is removed (a complete animation rule refers to a rule that can generate template codes in the order of "message type - applicable chart type - applicable animation object type - plot tree structure"); finally, the remaining candidate categories and candidate category combinations are merged according to the four aspects of message, scene, animation object and plot, respectively, to obtain the animation design space corresponding to the message dimension, scene dimension, animation object dimension and plot dimension.

[0030] Specifically, the animation design space includes message dimension, scene dimension, animation object dimension, and plot dimension. The message dimension is used to represent the type of data fact to be expressed; the scene dimension is used to represent the chart type of static infographics; the animation object dimension is used to represent data graphics, image elements, built-in text, and their semantic groupings; and the plot dimension is used to represent animation effects, rhythm, shots, and sequence.

[0031] The advantage of steps (2-2-1) to (2-2-8) above is that by statistically analyzing, counting the number of co-occurrences, and merging categories of multidimensional labels in the animation design samples, an animation design space of four dimensions—message, scene, animation object, and plot—can be automatically formed from the samples, reducing the subjectivity of manually defining animation categories.

[0032] (2-3) Establish a tree-structured set for the animation design space corresponding to the plot dimension obtained in step (2-2); such as Figure 3 As shown, the tree structure corresponding to the plot dimension is used to represent the tiling or nesting relationship between animation events. The root node represents the global rhythm type, the leaf node represents the smallest animation unit, and the internal rhythm node represents the organization method of the sub-animation units within the semantic group.

[0033] This step specifically includes the following sub-steps: (2-3-1) Set the first counter j=1; (2-3-2) Determine whether j is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-3-10); otherwise, proceed to step (2-3-3). (2-3-3) Obtain the animation process label of the j-th animation design sample in the animation design sample set, and use the global rhythm type in the animation process label as the root node of the plot tree corresponding to the j-th animation design sample; (2-3-4) Obtain the set of minimum animation units in the j-th animation design sample, and sort the set of minimum animation units according to the playback start time to obtain the sorted set of minimum animation units Uj={u1, u2, ..., uMj}, where Mj represents the total number of minimum animation units in the j-th animation design sample; (2-3-5) Set the second counter s=1; (2-3-6) Determine if s is greater than Mj. If so, proceed to step (2-3-9). Otherwise, read the s-th smallest animation unit us from the sorted set of smallest animation units and create a leaf node for the smallest animation unit us. The leaf node records the animation object type, animation effect, camera effect, start time and duration. (2-3-7) Determine whether the animation object type corresponding to the s-th smallest animation unit us is a semantic grouping type. If it is, obtain the semantic grouping object corresponding to the semantic grouping type and determine whether there are more than two sub-animation units inside the semantic grouping object. If not, keep the leaf node, set s=s+1, and return to step (2-3-6). If it exists, replace the leaf node with the internal rhythm node and go to step (2-3-8). Otherwise, keep the leaf node, set s=s+1, and return to step (2-3-6). (2-3-8) Obtain the set of sub-animation units inside the semantic grouping object, and sort the set of sub-animation units according to the playback start time; under the internal rhythm node, create leaf nodes corresponding to each sub-animation unit in sequence, and record the animation object type, animation effect, camera effect, start time and duration of the corresponding sub-animation unit in each leaf node; set s=s+1, and return to step (2-3-6). (2-3-9) Combine the root node, internal rhythm node, leaf node and child leaf node that have been established in the j-th animation design sample to form the plot tree corresponding to the animation design sample; set j=j+1 and return to step (2-3-2). (2-3-10) Perform isomorphic merging on the plot trees corresponding to all animation design samples. Merge plot trees with the same root node rhythm type, leaf node animation object type, leaf node animation effect, camera method and parent-child hierarchy into the same tree structure to obtain the set of tree structures corresponding to the plot dimension.

[0034] The advantage of steps (2-3-1) to (2-3-10) above is that by representing the plot dimension through a tree structure, it can represent both the tiling animation sequence in ordinary charts and the nested animation relationship within semantic groups in infographics.

[0035] (2-4) Based on the animation design space obtained in step (2-2) and the tree structure set obtained in step (2-3), an animation template library is established by using template encoding, rule merging and parameter statistics. This step specifically includes the following sub-steps: (2-4-1) Set the counter k=1; (2-4-2) Determine whether k is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-4-7); otherwise, proceed to step (2-4-3). (2-4-3) Read the message dimension label, scene dimension label, animation object dimension label and plot tree corresponding to the kth animation design sample; (2-4-4) The template encoding of the kth animation design sample is performed in the order of "message type - applicable chart type - applicable animation object type - plot tree structure" to obtain the template encoding result of the kth animation design sample; (2-4-5) Determine whether the template encoding result of the kth animation design sample obtained in step (2-4-4) already exists in the animation template library. If so, add the animation effect parameters, rhythm parameters and camera parameters of the kth animation design sample to the parameter set of the corresponding template; otherwise, create a new animation template record and write the template encoding result into the animation template record. (2-4-6) Set k=k+1 and return to step (2-4-2); (2-4-7) Perform statistical processing on the parameter set recorded in each animation template to obtain the default parameter value range of the duration, delay time, animation direction, scaling ratio and camera operation mode of the animation template; (2-4-8) Determine whether the number of times the sample appears in the animation template record is greater than or equal to the preset number threshold η. If so, retain the animation template record; otherwise, merge the animation template record into the animation template record that is most similar to its message type, applicable chart type and plot tree structure, or mark it as a low-frequency template; where the value of η is in the range of 2 to 20, preferably 3; (2-4-9) Assign a template number to each retained animation template record and generate an animation template library that includes message type, applicable chart type, applicable animation object type, animation effect, rhythm structure, shot mode, playback order and default parameter value range.

[0036] Each animation template in the animation template library includes at least the message type, applicable chart type, applicable animation object type, animation effect, rhythm structure, camera style, and playback order.

[0037] (2-5) Generate template matching information and animation mapping information based on the animation template library obtained in step (2-4); wherein, the template matching information is used to determine the matching score of the candidate animation template from the animation template library based on the comprehensive analysis results obtained in the online animation generation process; the animation mapping information is used to map the abstract animation objects, animation effects and playback order in the target animation template to the animation objects, animation parameters and executable animation logic in the static information chart; (2-6) Using the template matching information obtained in step (2-5), the comprehensive analysis result obtained in step (1) and the animation template library obtained in step (2-4) are matched and the candidate animation templates whose matching scores meet the preset conditions are determined as the target animation templates; when the number of candidate animation templates that meet the preset conditions is zero, the top K candidate animation templates with the highest matching scores are determined as the target animation templates in descending order of matching scores, so as to obtain multiple target animation templates; The advantage of steps (2-4) to (2-6) above is that, through template encoding, parameter statistics, template matching information and animation mapping information, the animation design space can be transformed into a searchable, matchable and executable mapping animation template library.

[0038] (3) Perform object mapping and plot parsing processing on all target animation templates obtained in step (2) to obtain executable animation logic; such as Figure 5 As shown, after obtaining the target animation template, this invention performs object mapping and plot parsing based on the plot tree structure, applicable animation object type set, animation effect set and camera mode in the template, thereby generating executable animation logic.

[0039] This step specifically includes the following sub-steps: (3-1) Set the counter r=1; (3-2) Determine whether r is greater than the total number R of target animation templates obtained in step (2). If so, proceed to step (3-9); otherwise, proceed to step (3-3). (3-3) Obtain the plot tree structure, applicable animation object type set, animation effect set, rhythm structure, camera mode and default parameter value range in the r-th target animation template; (3-4) For each leaf node in the plot tree structure, obtain the template animation object type corresponding to the leaf node, and filter all candidate animation objects with the same object type or that can be covered by the object type from the animation object group obtained in step (1); calculate the binding score Bind(o) of the candidate animation object based on the semantic relevance between each candidate animation object and the target data content, the spatial proximity between the candidate animation object and the target expression region, and the hierarchical consistency between the group to which the candidate animation object belongs and the template group, where Bind(o) = μ1·S r(o) + μ2·Sp(o) + μ3·Sh(o); Sr(o) represents the semantic relevance score between the candidate animation object o and the target data content, Sp(o) represents the spatial proximity score between the candidate animation object o and the target expression region, and Sh(o) represents the hierarchical consistency score between the group to which the candidate animation object o belongs and the template group; μ1, μ2 and μ3 are preset weights, the value of μ1 is from 0.1 to 1, preferably 0.4; the value of μ2 is from 0.1 to 1, preferably 0.4; the value of μ3 is from 0.1 to 1, preferably 0.2; (3-5) For each leaf node in the plot tree structure, determine whether the number of candidate animation objects corresponding to the leaf node obtained in step (3-4) is zero; if so, mark the leaf node as an unbound node and skip the unbound node when generating executable animation logic later; otherwise, select the candidate animation object with the largest binding score from all candidate animation objects corresponding to the leaf node and record its binding score as the maximum binding score; determine whether the maximum binding score is greater than or equal to the preset binding threshold β; if so, bind the candidate animation object corresponding to the maximum binding score to the specific animation object corresponding to the leaf node; otherwise, mark the leaf node as an unbound node and skip the unbound node when generating executable animation logic later; where β ranges from 0.3 to 1, preferably 0.6; after completing the judgment of all leaf nodes in the plot tree structure, obtain the leaf node binding result corresponding to the r-th target animation template, and proceed to step (3-6). (3-6) Based on the leaf node binding results obtained in step (3-5), the plot tree structure in the r-th target animation template is recursively traversed. When a bound leaf node is traversed, the animation event corresponding to that leaf node is generated. When an unbound node is traversed, the unbound node is skipped. When an internal rhythm node is traversed, the time relationship between its child nodes is determined according to the rhythm type of the internal rhythm node, thereby obtaining the animation event set and time relationship corresponding to the r-th target animation template. Specifically, the temporal relationship between the child nodes is determined according to the rhythm type of the internal rhythm node as follows: when the rhythm type is presented alone, an animation event is executed independently; when the rhythm type is presented in parallel, multiple animation events have the same start time; when the rhythm type is presented in stages, multiple animation events are executed sequentially; when the rhythm type is presented in staggered manner, multiple animation events are executed sequentially at preset time intervals; when the rhythm type is presented in layers, the next animation event starts before the previous animation event has completely finished.

[0040] (3-7) Based on the set of animation events and time relationships obtained in step (3-6), assign an animation start time and duration to each animation event, and generate corresponding animation effect parameters according to the animation effect type, thereby obtaining the set of animation event parameters corresponding to the r-th target animation template, including animation event, specific animation object, animation start time, duration, animation effect parameters and camera parameters; Specifically, the generated effects correspond to parameters such as transparency, scaling ratio, or displacement start point; the style change effects correspond to parameters such as color, transparency, stroke, or glow; the geometric change effects correspond to parameters such as position, rotation angle, scaling ratio, or shape change; the pointing effects correspond to the generation parameters of the arrow auxiliary element; the guiding effects correspond to the generation parameters of the path line auxiliary element; the explanation effects correspond to the generation parameters of the text box auxiliary element; and the highlight effects correspond to the generation parameters of the border auxiliary element. At the same time, scaling, translation, or toggling parameters are generated according to the camera mode. (3-8) Convert the set of animation event parameters corresponding to the r-th target animation template obtained in step (3-7) into the executable animation sub-logic corresponding to the target animation template, set r=r+1, and return to step (3-2). (3-9) Integrate all executable animation sub-logics corresponding to the target animation templates in descending order of template matching score to obtain executable animation logic; wherein, executable animation logic includes one or more of HTML structure, CSS animation style and JavaScript timeline control code.

[0041] The advantage of steps (3-1) to (3-9) above is that, through object binding, leaf node judgment, recursive traversal of the plot tree, and generation of animation event parameters, the target animation template can be converted into executable animation logic in a specific static information chart.

[0042] (4) Render the executable animation logic obtained in step (3) to obtain a dynamic chart animation corresponding to the static information chart.

[0043] Specifically, this step involves first determining the display state of each animation object on the timeline based on the executable animation logic obtained in step (3); then, dynamically rendering the data graphics, image elements, built-in text, or combinations thereof corresponding to the animation object in the static infographic based on the display state; and finally, synthesizing the dynamic rendering results corresponding to all animation objects to obtain a dynamic infographic animation that matches the user's input text.

[0044] The advantage of this step (4) is that it can dynamically render the data graphics, image elements, built-in text or combinations thereof in the static infographic according to the executable animation logic, thereby automatically generating dynamic infographic animations that match the user's input text.

[0045] Example 1: Construction of Animation Design Space and Animation Template Library In this embodiment, an animation design space for infographic animations is first constructed. The animation design space consists of message dimension, scene dimension, animation object dimension, and plot dimension.

[0046] The message dimension is used to represent the type of data fact that the animation aims to convey to the viewer. In a preferred embodiment, through systematic analysis of 1082 infographic animation clips from the real world, nine core message types were identified: Value (used to present precise numerical values ​​or proportions, accounting for 30.0% of the dataset), Trend (showing the process of data changing over time or in sequence, accounting for 12.6%), Difference (emphasizing the contrast between different objects, accounting for 12.5%), Extreme (highlighting the highest, lowest, or outlier values, accounting for 11.5%), Categorization (showing object grouping, accounting for 10.5%), Distribution (showing the range of data distribution, accounting for 8.9%), Rank (showing the ranking of objects, accounting for 8.3%), Association (showing the correlation between variables, accounting for 4.4%), and Outlier (highlighting outliers, accounting for 1.3%).

[0047] Scene dimensions are used to represent the type of visualization carrying the animation. In a preferred embodiment, scene dimensions include line charts (Line, 24.0%), bar charts (Bar, 22.6%), maps (Map, 11.5%), pie charts (Pie, 11.0%), timelines (Timeline, 10.0%), scatter plots (Scatter, 6.7%), bubble charts (Bubble, 6.3%), treemaps (Treemap, 3.0%), pyramids (Pyramid, 1.6%), Sankey diagrams (Sankey, 1.6%), gauge charts (Gauge, 1.5%), and matrix charts (Matrix, 0.4%).

[0048] The animation object dimension is used to represent the visual elements involved in the animation. Preferably, the animation object includes data graphics (used to encode data, such as bars, lines, and points), image elements (used for visual enhancement and metaphor, such as illustrations and icons), built-in text (used for numerical annotation and interpretation), and semantic groups (a collection of the above elements associated with the same data portion).

[0049] The plot dimension is used to represent the organization of animated events. Preferably, the plot dimension includes animation effects, rhythm, camera angles, and sequence. Animation effects can include generation (e.g., fade-in, pop-up, 31.4%), style changes (e.g., changing color, 12.1%), geometric changes (e.g., movement, scaling, 9.9%), disappearance (e.g., fade-out, 1.5%), highlighting (e.g., adding an outline, 9.6%), guiding (e.g., drawing a path, 12.7%), pointing (e.g., using arrows, 12.1%), explanation (e.g., adding a description box, 9.3%), and emphasis (e.g., a pulse effect, 0.7%). Rhythm can include individual presentation (40.6%), parallel presentation (19.8%), phased presentation (13.4%), staggered presentation (30.5%), and layered presentation (3.1%). Camera angles can include zooming (64.5%), panning (31.4%), and transitions (4.1%).

[0050] To describe the nested relationships in infographic animations, this embodiment uses a tree structure to represent the sequence in the plot dimension. The root node of the tree structure represents the global rhythm type (such as "phased presentation"). Leaf nodes represent a minimum animation unit, containing the animation object type, animation effect, and camera effect. When an animation object group contains sub-animation units, the leaf node corresponding to that group is replaced with a new rhythm node (internal node), and child leaf nodes representing the animation of its internal elements are set under that rhythm node. In this way, both 88.4% of the tiling animations in ordinary charts and 11.6% of the nested animation structures in infographics, with a maximum of three levels, can be represented.

[0051] After establishing the animation design space, an animation template library is constructed based on the combination relationships of the four dimensions. In a preferred embodiment, from the collected 1082 animations, 89 animation templates were summarized and refined through expert discussion and cross-validation for commonly used "scene-message" combinations (such as numerical messages in a pie chart, accounting for 58.9%). Each animation template includes a template number, applicable message type, applicable scene type, applicable animation object type, plot tree structure, and animation parameter rules (such as duration, delay, direction, etc.).

[0052] Example 2: Infographic Parsing and Semantic Grouping In this embodiment, the system receives a static infographic image input by the user. To generate an animation, the static image needs to be parsed into an operable animated object.

[0053] Fine-grained element parsing: The system first invokes a fine-grained object detection model. This model uses YOLO-v12 as its backbone network and is fine-tuned using a minimal animation unit-level annotation dataset containing 2,827 real-world infographics and a total of 169,896 bounding boxes. This dataset covers all 12 scenes and 9 message types. Testing showed that the detection model of this invention achieved an mAP of 0.7595@0.5 on the infographic character detection task, significantly outperforming InfoDet (0.1791) and YOLO-World (0.0572). After detection, the SAM model can be further applied, using the detection bounding boxes as cues, to obtain the precise segmentation mask for each element. The parsing results include the category (data graphics / image elements / embedded text), bounding box, and image region for each candidate animated object.

[0054] Semantic Grouping: After obtaining candidate animation objects, the system constructs an object relationship graph. Nodes in the graph represent candidate animation objects, and edge weights are calculated based on spatial proximity (i.e., characters that are close to each other are grouped together). The system clusters this graph to obtain semantic groups. For example, in a bar chart infographic, a bar, the numerical labels inside the bar, the character illustration next to the bar, and the category text below the bar can be grouped into the same semantic group. The grouping results are saved in a hierarchical structure, including group numbers, a list of internal animation objects, etc., providing a computable object basis for subsequent template matching.

[0055] Example 3: Text Semantic Parsing and Animation Template Matching In this embodiment, the system receives text input by the user, such as "UK total export freight value reached a low point in 2018". The system uses a large language model to parse the text, first determining that the data fact expressed by the text is of the type "extreme", and identifying the target object "UK total export freight value" and the time "2018".

[0056] After determining the message type, the system matches the message type with the semantic grouping obtained in Example 2 and locates the bar chart group (containing bars, icons and numerical labels) representing the "United Kingdom" data item.

[0057] The system then traverses the animation template library containing 89 templates, calculating a matching score for each candidate template. In a preferred embodiment, the matching score Score(T) is defined as: Score(T) = λ1·S_m(T) + λ2·S_v(T) + λ3·S_c(T), where T represents the candidate animation template, S_m(T) is the message matching degree (1 if the template message type matches the input message type, otherwise 0), S_v(T) is the scene matching degree (similarly), and S_c(T) is the animation object matching degree. The animation object matching degree S_c(T) further includes a support score (representing the frequency of the animation object combination in the template appearing in the library) and a utilization score (representing whether the animation object type required by the template is included in the parsed object type set). In this embodiment, the weights λ1, λ2, and λ3 can be set to 0.5, 0.5, and 1 respectively to prioritize the accuracy of animation object matching.

[0058] The system selects the highest-scoring template as the target animation template, or recommends the top 3 highest-scoring templates to the user for selection. In this way, the selection of the animation template is jointly determined by the text message, chart scene, and animation object.

[0059] Example 4: Mapping from Animation Template to Executable Animation Logic In this embodiment, after determining the target animation template, the system maps the target animation template to executable animation logic. Specifically, the system reads the plot tree structure in the target animation template and generates an animation event sequence according to the hierarchical order of the tree.

[0060] For each leaf node in the tree structure, the system binds the animation object type within the leaf node to the specific object or object group obtained in Example 2. For example, the root node of the template story tree is "Phased Presentation," which has three child nodes corresponding to three data groups (e.g., China, the United States, and the United Kingdom). The system will play animations for these three groups sequentially. Within each group, if the story tree defines it as an internal rhythm node (e.g., "Staggered Presentation"), the system will assign different delay times to its internal columns, icons, and text to produce a staggered animation effect.

[0061] The system calculates parameters for each animated object based on template rules. If the animation effect is "Generate," the transparency is scaled from 0 to 1 or from small to large; if it's "Style Change," the target color is set; if it's "Highlight," the corresponding border guidelines are generated. Finally, the system converts the animation event sequence into HTML, CSS, and JavaScript code, outputting an interactive web animation that fully preserves the visual style of the original infographic.

[0062] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for automatically generating infographic animations based on animation design space, characterized in that, Includes the following steps: (1) Obtain the user input text and the static infographic paired with the user input text, and parse the user input text and the static infographic to obtain a comprehensive analysis result; (2) Using the pre-established animation design space and animation template library, perform template matching processing on the comprehensive analysis results obtained in step (1) to obtain multiple target animation templates; (3) Perform object mapping and plot parsing processing on all target animation templates obtained in step (2) to obtain executable animation logic; (4) Render the executable animation logic obtained in step (3) to obtain a dynamic chart animation corresponding to the static information chart.

2. The method for automatically generating infographic animations based on animation design space according to claim 1, characterized in that, Step (1) specifically involves first obtaining the user input text and the static infographic paired with the user input text; then, performing semantic parsing on the user input text to extract the message type, target data content, and target expression intent from the user input text. Subsequently, the static infographics are subjected to standardization preprocessing to obtain standardized preprocessed static infographics. Subsequently, fine-grained element parsing is performed on the standardized preprocessed static infographics to obtain multiple data graphics, image elements, and embedded text. Then, the obtained data graphics, image elements, and embedded text are semantically grouped to obtain animation object groups. Semantic grouping refers to grouping data graphics, image elements, and embedded text that correspond to the same data item, the same visual area, or the same expressive meaning into the same animation object group. After fine-grained element parsing, the static infographics yield multiple candidate animation objects, and animation object groups are formed based on the spatial proximity and semantic relationship between objects. Finally, the obtained message type, target data content, target expressive intent, and animation object groups are combined into a comprehensive analysis result.

3. The method for automatically generating infographic animations based on animation design space according to claim 1 or 2, characterized in that, Step (2) specifically involves first selecting multiple candidate animation templates from a pre-established animation template library based on the message type and the chart type of the static infographic in the comprehensive analysis result; then, representing the comprehensive analysis result as an input triple I={m, v, e}, where m represents the message type corresponding to the user input text, v represents the scene type corresponding to the static infographic, and e represents the set of animation object types obtained after fine-grained element analysis and semantic grouping. The set of animation object types e includes one or more of data graphics, image elements, and built-in text. Subsequently, for each candidate animation template t, its template triple {mt, vt, et} is obtained, where mt represents the template message type, vt represents the applicable scenario type of the template, and et represents the set of animation object types required by the template. Then, the message matching score Sm(t, m), scenario matching score Sv(t, v), and animation object matching score Se(t, e) of the candidate animation template t are calculated respectively, and the matching score of the candidate animation template t is calculated according to Score(t) = λ1·Sm(t, m) + λ2·Sv(t, v) + λ3·Se(t, e). Finally, the candidate animation templates whose matching scores meet the preset conditions are determined as the target animation templates. Where λ1, λ2, and λ3 are preset weights; Wherein, when mt is the same as m, Sm(t, m) = 1, otherwise Sm(t, m) = 0; when vt is the same as v, Sv(t, v) = 1, otherwise Sv(t, v) = 0; the animation object matching score Se(t, e) is determined by the support score and the utilization score. The support score is used to represent the frequency of occurrence of the set of animation object types et required by the template in the animation template library, and the utilization score is used to represent the degree to which the set of animation object types et required by the template is covered by the set of input animation object types e; when et cannot be covered by e, Se(t, e) = 0, when et can be covered by e, Se(t, e) is equal to the sum of the support score and the utilization score; The preset conditions are: the matching score of the candidate animation template is greater than or equal to the preset matching threshold θ, the message type of the candidate animation template is consistent with the message type in the comprehensive analysis result, the applicable chart type of the candidate animation template is consistent with the chart type of the static information chart, and the applicable animation object type of the candidate animation template is covered by any object type in the animation object group; where θ ranges from 0.5 to 3. When the number of candidate animation templates that meet the above preset conditions is zero, the top K candidate animation templates with the highest matching scores are determined as the target animation templates, where K is a positive integer.

4. The method for automatically generating infographic animations based on animation design space according to any one of claims 1 to 3, characterized in that, The animation design space and animation template library in step (2) are established according to the following steps: (2-1) Obtain multiple infographic animation samples, and annotate the text content, chart type, animation object and animation process in all infographic animation samples. The annotated text content, chart type, animation object and animation process constitute the animation design sample set. (2-2) Perform multidimensional label statistics, co-occurrence count statistics and category merging on the animation design sample set obtained in step (2-1) to obtain an animation design space composed of multiple dimensions; The animation design space includes message dimension, scene dimension, animation object dimension, and plot dimension. These dimensions work together to describe the data messages, applicable scenarios, objects involved in the animation, and the way the animation plot is organized in the infographic animation. (2-3) Establish a tree structure set for the animation design space corresponding to the plot dimension obtained in step (2-2); the tree structure corresponding to the plot dimension is used to represent the tiling relationship or nesting relationship between animation events, where the root node represents the global rhythm type, the leaf node represents the smallest animation unit, and the internal rhythm node represents the organization method of the sub-animation units within the semantic group. (2-4) Based on the animation design space obtained in step (2-2) and the tree structure set obtained in step (2-3), an animation template library is established by using template encoding, rule merging and parameter statistics. (2-5) Generate template matching information and animation mapping information based on the animation template library obtained in step (2-4); wherein, the template matching information is used to determine the matching score of the candidate animation template from the animation template library based on the comprehensive analysis results obtained in the online animation generation process; the animation mapping information is used to map the abstract animation objects, animation effects and playback order in the target animation template to the animation objects, animation parameters and executable animation logic in the static information chart; (2-6) Using the template matching information obtained in step (2-5), the comprehensive analysis result obtained in step (1) and the animation template library obtained in step (2-4) are matched and the candidate animation templates whose matching scores meet the preset conditions are determined as the target animation templates. When the number of candidate animation templates that meet the preset conditions is zero, the top K candidate animation templates with the highest matching scores are determined as the target animation templates in descending order of matching scores, so as to obtain multiple target animation templates.

5. The method for automatically generating infographic animations based on animation design space according to claim 4, characterized in that, Step (2-2) specifically includes the following sub-steps: (2-2-1) Set the counter i=1 and initialize the global standard label set to empty; (2-2-2) Determine whether i is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-2-6); otherwise, proceed to step (2-2-3). (2-2-3) Obtain the message tag, scene tag, animation object tag, and animation process tag of the i-th animation design sample in the animation design sample set; Specifically, message tags are used to characterize the type of data fact conveyed by the animation design sample, scene tags are used to characterize the type of infographic corresponding to the animation design sample, animation object tags are used to characterize the data graphics, image elements, built-in text or semantic groups involved in the animation, and animation process tags are used to characterize the animation effects, rhythm, shots and sequence; among them, rhythm includes one or more of the following: individual presentation, parallel presentation, phased presentation, staggered presentation and layered presentation. (2-2-4) Normalize and merge the tags with the same meaning or the same function in the i-th animation design sample in the animation design sample set to obtain the standard tag set corresponding to the animation design sample, including message standard tag subset, scene standard tag subset, animation object standard tag subset, and animation process standard tag subset; (2-2-5) Add the set of standard labels corresponding to the i-th animation design sample obtained in step (2-2-4) to the global standard label set, set i=i+1, and return to step (2-2-2). (2-2-6) Count the frequency of occurrence of each standard label in the global standard label set, and count the number of times each pair of standard labels co-occurs in the same animation design sample; (2-2-7) Determine the candidate categories and candidate category combinations based on the frequency of occurrence and co-occurrence of the standard labels as calculated in step (2-2-6); Specifically, when the frequency of a standard tag is greater than or equal to the first frequency threshold α, the standard tag is identified as a candidate category, and the dimension to which the candidate category belongs is recorded according to the message standard tag subset, scene standard tag subset, animation object standard tag subset, or animation process standard tag subset to which the standard tag belongs; where, when the standard tag belongs to the animation process standard tag subset, its dimension is recorded as the plot dimension; the value range of the first frequency threshold α is 2 to 30. When the co-occurrence frequency of two standard tags is greater than or equal to the second frequency threshold γ, or the co-occurrence ratio of two standard tags is greater than or equal to the co-occurrence ratio threshold ρ, the two standard tags are set as a candidate category combination that satisfies the co-occurrence condition; where, the value range of the second frequency threshold γ is 2 to 20; the co-occurrence ratio is the ratio between the co-occurrence frequency of two standard tags and the smaller value of the occurrence frequency of the two standard tags, and the value range of the co-occurrence ratio threshold ρ is 0.1 to 0.

8. (2-2-8) Synonym tags are merged for the candidate categories and candidate category combinations obtained in step (2-2-7), and the candidate categories and candidate category combinations are filtered according to the number of co-occurrences, the co-occurrence ratio and the frequency of category occurrence to obtain the animation design space corresponding to the message dimension, scene dimension, animation object dimension and plot dimension. Specifically, this step involves first processing each candidate category according to its dimension, without merging candidate categories from different dimensions; then, merging candidate categories with the same meaning or function within the same dimension to obtain higher-level candidate categories; subsequently, screening candidate category combinations based on their co-occurrence frequency and co-occurrence ratio. When the co-occurrence frequency of two candidate categories is greater than or equal to a second frequency threshold γ, or the co-occurrence ratio of two candidate categories is greater than or equal to a co-occurrence ratio threshold ρ, the candidate category combination is retained. The second frequency threshold γ ranges from 2 to 20; the co-occurrence ratio is the smaller of the co-occurrence frequency and the frequency of occurrence of the two candidate categories. The ratio between values, the co-occurrence ratio threshold ρ ranges from 0.1 to 0.8; when the frequency of a candidate category is less than the first frequency threshold α, and the candidate category cannot form a complete animation rule with other candidate categories, the candidate category is removed; the first frequency threshold α ranges from 2 to 30; when a candidate category combination does not meet the above co-occurrence conditions and cannot form a complete animation rule, the candidate category combination is removed; finally, the remaining candidate categories and candidate category combinations are merged according to four aspects: message, scene, animation object, and plot, to obtain the animation design space corresponding to the message dimension, scene dimension, animation object dimension, and plot dimension, respectively; Specifically, the animation design space includes message dimension, scene dimension, animation object dimension, and plot dimension; message dimension is used to represent the type of data fact to be expressed; scene dimension is used to represent the chart type of static infographics; animation object dimension is used to represent data graphics, image elements, built-in text and their semantic groupings; plot dimension is used to represent animation effects, rhythm, shots and sequence.

6. The method for automatically generating infographic animations based on animation design space according to claim 5, characterized in that, Step (2-3) specifically includes the following sub-steps: (2-3-1) Set the first counter j=1; (2-3-2) Determine whether j is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-3-10); otherwise, proceed to step (2-3-3). (2-3-3) Obtain the animation process label of the j-th animation design sample in the animation design sample set, and use the global rhythm type in the animation process label as the root node of the plot tree corresponding to the j-th animation design sample; (2-3-4) Obtain the set of minimum animation units in the j-th animation design sample, and sort the set of minimum animation units according to the playback start time to obtain the sorted set of minimum animation units Uj={u1, u2, ..., uMj}, where Mj represents the total number of minimum animation units in the j-th animation design sample; (2-3-5) Set the second counter s=1; (2-3-6) Determine if s is greater than Mj. If so, proceed to step (2-3-9). Otherwise, read the s-th smallest animation unit us from the sorted set of smallest animation units and create a leaf node for the smallest animation unit us. The leaf node records the animation object type, animation effect, camera effect, start time and duration. (2-3-7) Determine whether the animation object type corresponding to the s-th smallest animation unit us is a semantic grouping type. If it is, obtain the semantic grouping object corresponding to the semantic grouping type and determine whether there are more than two sub-animation units inside the semantic grouping object. If not, keep the leaf node, set s=s+1, and return to step (2-3-6). If it exists, replace the leaf node with the internal rhythm node and go to step (2-3-8). Otherwise, keep the leaf node, set s=s+1, and return to step (2-3-6). (2-3-8) Obtain the set of sub-animation units inside the semantic grouping object, and sort the set of sub-animation units according to the playback start time; under the internal rhythm node, create leaf nodes corresponding to each sub-animation unit in sequence, and record the animation object type, animation effect, camera effect, start time and duration of the corresponding sub-animation unit in each leaf node; set s=s+1, and return to step (2-3-6). (2-3-9) Combine the root node, internal rhythm node, leaf node and child leaf node that have been established in the j-th animation design sample to form the plot tree corresponding to the animation design sample; Set j = j + 1 and return to step (2-3-2); (2-3-10) Perform isomorphic merging on the plot trees corresponding to all animation design samples. Merge plot trees with the same root node rhythm type, leaf node animation object type, leaf node animation effect, camera method and parent-child hierarchy into the same tree structure to obtain the set of tree structures corresponding to the plot dimension.

7. The method for automatically generating infographic animations based on animation design space according to claim 6, characterized in that, Steps (2-4) specifically include the following sub-steps: (2-4-1) Set the counter k=1; (2-4-2) Determine whether k is greater than the total number N of animation design samples in the animation design sample set. If yes, proceed to step (2-4-7); otherwise, proceed to step (2-4-3). (2-4-3) Read the message dimension label, scene dimension label, animation object dimension label and plot tree corresponding to the kth animation design sample; (2-4-4) The template encoding of the kth animation design sample is performed in the order of "message type - applicable chart type - applicable animation object type - plot tree structure" to obtain the template encoding result of the kth animation design sample; (2-4-5) Determine whether the template encoding result of the kth animation design sample obtained in step (2-4-4) already exists in the animation template library. If so, add the animation effect parameters, rhythm parameters and camera parameters of the kth animation design sample to the parameter set of the corresponding template; otherwise, create a new animation template record and write the template encoding result into the animation template record. (2-4-6) Set k=k+1 and return to step (2-4-2); (2-4-7) Perform statistical processing on the parameter set recorded in each animation template to obtain the default parameter value range of the duration, delay time, animation direction, scaling ratio and camera operation mode of the animation template; (2-4-8) Determine whether the number of times the sample appears in the animation template record is greater than or equal to the preset number threshold η. If so, retain the animation template record; otherwise, merge the animation template record into the animation template record that is most similar to its message type, applicable chart type and plot tree structure, or mark it as a low-frequency template; where the value of η ranges from 2 to 20. (2-4-9) Assign a template number to each retained animation template record and generate an animation template library containing message type, applicable chart type, applicable animation object type, animation effect, rhythm structure, shot mode, playback order and default parameter value range; Each animation template in the animation template library includes at least the message type, applicable chart type, applicable animation object type, animation effect, rhythm structure, camera style, and playback order.

8. The method for automatically generating infographic animations based on animation design space according to claim 7, characterized in that, Step (3) specifically includes the following sub-steps: (3-1) Set the counter r=1; (3-2) Determine whether r is greater than the total number R of target animation templates obtained in step (2). If so, proceed to step (3-9); otherwise, proceed to step (3-3). (3-3) Obtain the plot tree structure, applicable animation object type set, animation effect set, rhythm structure, camera mode and default parameter value range in the r-th target animation template; (3-4) For each leaf node in the plot tree structure, obtain the template animation object type corresponding to the leaf node, and filter all candidate animation objects with the same object type or that can be covered by the object type from the animation object group obtained in step (1); calculate the binding score Bind(o) of the candidate animation object based on the semantic relevance between each candidate animation object and the target data content, the spatial proximity between the candidate animation object and the target expression region, and the hierarchical consistency between the group to which the candidate animation object belongs and the template group, where Bind(o) o) = μ1·Sr(o) + μ2·Sp(o) + μ3·Sh(o); Sr(o) represents the semantic relevance score between the candidate animation object o and the target data content, Sp(o) represents the spatial proximity score between the candidate animation object o and the target expression region, and Sh(o) represents the hierarchical consistency score between the group to which the candidate animation object o belongs and the template group; μ1, μ2, and μ3 are preset weights, with μ1 ranging from 0.1 to 1; μ2 ranging from 0.1 to 1; and μ3 ranging from 0.1 to 1, preferably 0.2; (3-5) For each leaf node in the plot tree structure, determine whether the number of candidate animation objects corresponding to the leaf node obtained in step (3-4) is zero; if so, mark the leaf node as an unbound node and skip the unbound node when generating executable animation logic in the future; otherwise, select the candidate animation object with the largest binding score from all candidate animation objects corresponding to the leaf node and record its binding score as the maximum binding score; determine whether the maximum binding score is greater than or equal to the preset binding threshold β; if so, bind the candidate animation object corresponding to the maximum binding score to the specific animation object corresponding to the leaf node; otherwise, mark the leaf node as an unbound node and skip the unbound node when generating executable animation logic in the future; where β ranges from 0.3 to 1; after completing the judgment of all leaf nodes in the plot tree structure, obtain the leaf node binding result corresponding to the r-th target animation template and proceed to step (3-6). (3-6) Based on the leaf node binding results obtained in step (3-5), the plot tree structure in the r-th target animation template is recursively traversed. When a bound leaf node is traversed, the animation event corresponding to that leaf node is generated. When an unbound node is traversed, the unbound node is skipped. When an internal rhythm node is traversed, the time relationship between its child nodes is determined according to the rhythm type of the internal rhythm node, thereby obtaining the animation event set and time relationship corresponding to the r-th target animation template. (3-7) Based on the set of animation events and time relationships obtained in step (3-6), assign an animation start time and duration to each animation event, and generate corresponding animation effect parameters according to the animation effect type, thereby obtaining the set of animation event parameters corresponding to the r-th target animation template, including animation event, specific animation object, animation start time, duration, animation effect parameters and camera parameters; (3-8) Convert the set of animation event parameters corresponding to the r-th target animation template obtained in step (3-7) into the executable animation sub-logic corresponding to the target animation template, set r=r+1, and return to step (3-2). (3-9) Integrate all executable animation sub-logics corresponding to the target animation templates in descending order of template matching score to obtain executable animation logic; wherein, executable animation logic includes one or more of HTML structure, CSS animation style and JavaScript timeline control code.

9. The method for automatically generating infographic animations based on animation design space according to claim 8, characterized in that, In steps (3-6), determining the temporal relationship between child nodes based on the rhythm type of the internal rhythm node specifically involves the following: when the rhythm type is single presentation, an animation event is executed independently; when the rhythm type is parallel presentation, multiple animation events have the same start time; when the rhythm type is staged presentation, multiple animation events are executed sequentially; when the rhythm type is staggered presentation, multiple animation events are executed sequentially at preset time intervals; and when the rhythm type is layered presentation, the next animation event starts before the previous animation event has completely finished. In steps (3-7), the generated effects correspond to the parameters of transparency, scaling ratio, or displacement start point; the style change effects correspond to the parameters of color, transparency, stroke, or glow; the geometric change effects correspond to the parameters of position, rotation angle, scaling ratio, or shape change; the pointing effects correspond to the generation parameters of the arrow auxiliary element; the guiding effects correspond to the generation parameters of the path line auxiliary element; the explanation effects correspond to the generation parameters of the text box auxiliary element; and the highlight effects correspond to the generation parameters of the border auxiliary element. At the same time, scaling, translation, or toggling parameters are generated according to the camera mode. Step (4) specifically involves first determining the display state of each animation object on the timeline based on the executable animation logic obtained in step (3); then, dynamically rendering the data graphics, image elements, built-in text, or combinations thereof corresponding to the animation object in the static infographic based on the display state; and finally, synthesizing the dynamic rendering results corresponding to all animation objects to obtain a dynamic infographic animation that matches the user's input text.

10. An automatic infographic animation generation system based on animation design space, characterized in that, Includes the following modules: The first module is used to obtain user input text and static infographics paired with user input text, and to parse user input text and static infographics to obtain comprehensive analysis results; The second module is used to perform template matching processing on the comprehensive analysis results obtained from the first module using a pre-established animation design space and animation template library to obtain multiple target animation templates. The third module is used to perform object mapping and plot parsing on all the target animation templates obtained from the second module in order to obtain executable animation logic; The fourth module is used to render the executable animation logic obtained from the third module to obtain a dynamic chart animation corresponding to the static infographic.

Images