Textual description-based automatic generation method for trajectory narrative visualization

Abstract

The present invention relates to a textual description-based automatic generation method for trajectory narrative visualization, comprising: using a large language model to extract structured trajectory description information from natural language descriptions to be analyzed, said natural language descriptions comprising textual descriptions in a natural language form and structured trajectory records; matching the extracted structured trajectory description information and the trajectory records, so as to form a series of story scenes in visualization; in accordance with a preset visual encoding rule, visually rendering the formed story scenes, so as to render same into a focus visualization layer and an environmental visualization layer. The method disclosed by the present inventio can automatically generate visualization of a single trajectory narrative on the basis of natural language textual descriptions and trajectory records, helps lowering the user threshold for trajectory narrative visualization, and supports expanding the use of narrative visualization in scenarios such as education, exhibition and media, thus laying a reliable foundation for the development of trajectory narrative visualization.

Description

A method for automatic generation of trajectory story narrative visualization based on text description Technical Field

[0001] The present invention belongs to the field of visualization, and in particular relates to a method for automatically generating trajectory story narration visualization based on text description. Background Art

[0002] A single spatiotemporal trajectory is crucial for analyzing and understanding the backgrounds and life experiences of historical figures. Among the many methods that can facilitate user understanding of historical figures, narrative visualization can intuitively display the movement of a single trajectory, further enhancing user understanding of historical figures and achieving outstanding results in educational and science applications. However, creating narrative visualizations typically requires proficiency in specialized tools and is a tedious task. Existing tools and methods lack support for narrators to easily create personalized trajectory stories.

[0003] Summary of the Invention

[0004] To address the shortcomings of the existing technology, the present invention aims to provide a method for automatically generating visualizations of trajectory narratives based on text descriptions. This method allows users to easily and flexibly create visualizations of single trajectory narratives, reducing the burden on users and lowering the technical requirements for creation while also accommodating the diverse creative needs of narrative creators. Users can input a natural language trajectory narrative and the corresponding trajectory record data, and a visualization is automatically generated. This visualization is designed to both align with the user's natural language story script and faithfully reflect the trajectory record data.

[0005] To achieve the above objectives, the present invention adopts a technical solution: a method for automatically generating a visual trajectory story narrative based on text description, the method comprising the following steps:

[0006] S1. Extract structured trajectory description information from the natural language description to be analyzed using a large language model;

[0007] S2. Match the extracted structured trajectory description information with the trajectory records in the natural language description to be analyzed to form a series of story scenes in the visualization;

[0008] S3. Comply with the preset visual encoding rules and perform visual rendering on the formed story scene.

[0009] Furthermore, the natural language description to be analyzed in step S1 includes a text description in natural language form and a structured trajectory record.

[0010] Furthermore, the track record includes the narrative structure of the entire track, and the text description reflects the user's emphasis on story combination and narrative focus.

[0011] Furthermore, the large language model in step S1 converts the structured trajectory description information into a structured trajectory descriptor, where the trajectory descriptor includes a time description and a location description.

[0012] Furthermore, the trajectory descriptor includes not only direct time information and location information, but also correlations between the time information and location information.

[0013] Furthermore, step S2 includes combining the time information and location information in the trajectory descriptor into an event descriptor containing spatiotemporal information, and matching the event descriptor to the trajectory record with the help of external knowledge and fuzzy search, ultimately forming a series of story scenes in the visualization.

[0014] Furthermore, the preset visual encoding rules in step S3 include a focus visualization layer and an environment visualization layer.

[0015] Furthermore, the focus visualization layer displays the user's direct narrative objects, and the environment visualization layer displays related trajectory records to complement and support the user's direct narrative.

[0016] Furthermore, in the focus visualization layer, the stay and movement events directly mentioned by the user in a sentence are directly drawn on the map using circles and arc arrows.

[0017] Furthermore, in the environment visualization layer, relevant trajectory records are displayed in a semi-transparent manner, and the trajectory records show the connection between the key stay and movement events directly narrated by the user, or show the missing elements in the user's description.

[0018] The beneficial technical effect of the present invention is that the method for automatically generating a trajectory story narrative visualization based on text description disclosed in the present invention can automatically generate a single trajectory story narrative visualization from a natural language text description and trajectory records, helping users to flexibly create a single trajectory story narrative visualization with low threshold, and supporting the expansion of the application of narrative visualization in scenarios such as education, exhibitions, and media. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] FIG1 is a flow chart of a method for automatically generating a visual trajectory story narrative based on text description according to a first embodiment of the present invention;

[0020] FIG2 is a schematic diagram of the contents of the natural language description to be analyzed input by the user in the method for automatically generating a trajectory story narrative visualization based on text description according to the first embodiment of the present invention;

[0021] FIG3 is a schematic diagram of a preset visual coding rule in a method for automatically generating a trajectory story narrative visualization based on text description according to the first embodiment of the present invention;

[0022] FIG4 is a schematic diagram of a scenario of a trajectory story narration and its corresponding design space generated by a method for automatically generating a trajectory story narration visualization based on text description according to the first embodiment of the present invention;

[0023] FIG5 shows a visualization result of the trajectory story narration of the poet Li Bai generated by the method for automatically generating trajectory story narration visualization based on text description according to the first embodiment of the present invention. DETAILED DESCRIPTION

[0024] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0025] Example 1

[0026] An embodiment of the present invention provides a method for automatically generating trajectory story narrative visualization based on text descriptions. By analyzing user needs and data task characteristics, the flow of information in the task is abstracted. The process of creating a single-person trajectory narrative visualization is abstracted into the matching and synthesis of data facts from text descriptions and story structures and emphasized information from trajectory records.

[0027] As shown in FIG1 , an embodiment of the present invention provides a method for automatically generating a visual trajectory story narrative based on text description, the method comprising the following steps:

[0028] S1. Using a large language model, structured trajectory description information is extracted from the natural language description to be analyzed. The structured trajectory description information is represented in the form of trajectory descriptors.

[0029] When creating data-driven storytelling, a carefully designed narrative structure and focus are crucial, while also staying true to the underlying data. As shown in Figure 2, to meet these two requirements, in this embodiment of the present invention, the user-entered natural language description to be analyzed includes a natural language text description and a structured trajectory log. These two components serve as the primary sources of data facts and narrative structure, respectively, ultimately synthesizing the trajectory storytelling visualization. The trajectory log encompasses the entire trajectory, while the text description reflects the user's emphasis on story composition and narrative focus.

[0030] To address the complexity of natural language presentation of trajectory descriptions, a large language model is used to understand trajectory descriptions. By setting prompt words to indicate the task and providing examples in the prompt words, the large language model converts the trajectory description into a structured trajectory descriptor. The descriptor includes time description and location description. In addition to direct time and location information, the descriptor also includes the correlation between time and location. The correlation with time includes just before (<=), after (>=), and exactly (=). The correlation with location includes leaving (from), arriving (to), and staying (stay).

[0031] S2. Match the extracted structured trajectory description information with the trajectory records to form a series of story scenes in the visualization.

[0032] Based on the extracted trajectory descriptor sequence, a trajectory algorithm is used to match the trajectory descriptor sequence with the trajectory record. First, the time and location information in the trajectory descriptor are combined into an event descriptor containing spatiotemporal information. Then, external knowledge (such as a geographic administrative division tree) and fuzzy search are used to match it to the trajectory record, ultimately forming a series of story scenes in the visualization.

[0033] S3. Comply with the preset visual encoding rules and perform visual rendering on the formed story scene.

[0034] The story scenes formed after matching are converted into visual elements and rendered into story narrative visualization. Following the preset visual encoding rules, the user's direct description is matched with the relevant trajectory records and rendered into key visualization layers and environmental visualization layers.

[0035] As shown in Figure 3, to better reflect the narrative structure and data facts in trajectory storytelling visualization, and to highlight the user's direct narrative intent, the preset visual encoding rules include a highlight visualization layer and a context visualization layer. The highlight visualization layer displays the user's direct narrative object, while the context visualization layer displays related trajectory records, completing and supporting the user's direct narrative.

[0036] As shown in Figure 4, in the focus visualization layer, circles and arc arrows are used to directly draw the stay and movement events directly mentioned by the user in a sentence on the map, while in the environment visualization layer, related trajectory segments are displayed in a semi-transparent manner. These related trajectory segments show the connection between the key stay and movement events directly narrated by the user, or show the missing elements in the user's description (such as the missing starting point in the narration of "leaving a place").

[0037] Two visualization layers are used to highlight the user's direct narrative content and the underlying trajectory information.

[0038] As shown in Figure 5, a method for automatically generating trajectory storytelling visualizations based on text descriptions disclosed in this invention was used to generate a visualization of the poet Li Bai's trajectory. The image shows two representative scenes, demonstrating how narrative text, with emphasis on regions and movement, can generate narrative trajectory visualizations.

[0039] The above embodiments show that the method for automatically generating trajectory story narrative visualization based on text description disclosed in the present invention can automatically generate a single trajectory story narrative visualization based on natural language text description and trajectory records, helping users to lower the threshold for trajectory story narrative visualization, supporting the expansion of narrative visualization applications in scenarios such as education, exhibitions, and media, and laying a solid foundation for the development of trajectory story narrative visualization.

[0040] The method described in the present invention is not limited to the embodiments described in the specific implementation manner. Those skilled in the art may derive other implementation manners based on the technical solution of the present invention, which also fall within the scope of the technical innovation of the present invention.

Claims

1. An automatic generation method for visualizing trajectory story narration based on text description, the method comprising the following steps: S1. Extract structured trajectory description information from the natural language description to be analyzed using a large language model; S2. Match the extracted structured trajectory description information with the trajectory records in the natural language description to be analyzed to form story scenes in visualization; S3. Render the formed story scenes visually in accordance with the preset visual coding rules.

2. The automatic generation method for visualizing trajectory story narration based on text description according to claim 1, characterized in that: In step S1, the natural language description to be analyzed includes text descriptions in natural language form and structured trajectory records.

3. The automatic generation method for visualizing trajectory story narration based on text description according to claim 2, characterized in that: The trajectory record contains the narrative structure of the entire trajectory, and the text description reflects the user's emphasis on story combination and narrative focus.

4. The automatic generation method for visualizing trajectory story narration based on text description according to claim 3, characterized in that: In step S1, the large language model converts the structured trajectory description information into structured trajectory description sub-elements, and the trajectory description sub-elements include time description and location description.

5. The automatic generation method for visualizing trajectory story narration based on text description according to claim 4, characterized in that: In addition to the direct time information and location information, the trajectory description sub-elements also include the correlation relationships with the time information and location information.

6. The automatic generation method for visualizing trajectory story narration based on text description according to claim 5, characterized in that: Step S2 includes combining the time information and location information in the trajectory description sub-elements into event description sub-elements containing spatio-temporal information, matching the event description sub-elements to the trajectory records with the help of external knowledge and fuzzy search, and finally forming story scenes in visualization.

7. The automatic generation method for visualizing trajectory story narration based on text description according to claim 6, characterized in that: The preset visual coding rules in step S3 include a key visualization layer and an environmental visualization layer.

8. The automatic generation method for visualizing trajectory story narration based on text description according to claim 7, characterized in that: The key visualization layer shows the user's direct narrative object, and the environmental visualization layer shows the relevant trajectory records to complement and support the user's direct narrative.

9. The automatic generation method for visualizing trajectory story narration based on text description according to claim 8, characterized in that: In the key visualization layer, circles and arc arrows are directly drawn on the map to represent the stay and movement events directly mentioned by the user in a sentence.

10. The automatic generation method for visualizing trajectory story narration based on text description according to claim 8, characterized in that: In the environmental visualization layer, the relevant trajectory records are shown in a semi-transparent manner, and the trajectory records show the connections between the key stay and movement events directly narrated by the user, or show the elements missing from the user's description.

Images