Data Visualization Query Answering via Feature Map Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computer search techniques are unable to access or utilize the underlying data in data visualizations, leading to users missing relevant information embedded within image files, as these techniques primarily focus on text searches and ignore visual data.
Innovation Solution
A computer program product that identifies data visualizations, generates a feature map characterizing spatial relationships, encodes queries as feature vectors, and predicts answer locations within the visualization, enabling the extraction of information from data visualizations, including those embedded as image files.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional text-based search techniques are used, then search processing is simple and fast, but the ability to access information from data visualizations is lost
Solution Approach 1:
The patent introduces an intermediary system consisting of a feature map generator and query feature vector generator that bridges the gap between text queries and visual data. The feature map generator converts data visualization images into feature maps, while the query feature vector generator transforms text queries into feature vectors. These intermediaries enable the system to process both visual and textual information, allowing users to search for information within data visualizations while maintaining search efficiency.
2Ease of manufacture
If data visualizations are treated as image files, then they can be stored and displayed, but computer search techniques cannot access the underlying data
Solution Approach 1:
The patent extracts the underlying data from the visual representation by generating feature maps that capture spatial relationships and data characteristics. The feature map generator processes the data visualization image and extracts meaningful features, while the query feature vector generator identifies relevant information based on user queries. This extraction process makes the underlying data accessible to search techniques while preserving the visual nature of the data visualization for storage and display.
3Productivity
If conventional search techniques are applied to documents with embedded visualizations, then text search is efficient, but queries cannot be satisfied by visualization-containing documents
Solution Approach 1:
The patent creates a universal search system that can handle both text-based queries and visual data queries. The feature map generator processes data visualizations into feature maps that can be queried alongside text documents. The query feature vector generator can process both text queries and visualize query requirements. This multi-functional approach allows the search system to satisfy queries regardless of whether the information is contained in text or visualizations, while maintaining the efficiency of conventional search techniques.
Data Source
AI summary
Systems and techniques are described that provide for question answering using data visualizations, such as bar graphs. Such data visualizations are often generated from collected data, and provided within image files that illustrate the underlying data and relationships between data elements. The described techniques analyze a query and a related data visualization, and identify one or more spatial regions within the data visualization in which an answer to the query may be found.