Vector Graphic Text Block Semantic Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vector graphic format documents lack logical structures, making it difficult to reuse and repurpose digital content, as they only maintain visual appearance and not the structural information like paragraphs or titles.
Innovation Solution
A computer-implemented method that analyzes the content of a vector graphic format document by dividing text into blocks, pre-classifying them, and combining them into a text flow with defined semantic roles, using statistical models based on typical font, row length, and deviation to recognize body text and non-body text blocks, and rearranging them into a semantically arranged data structure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If vector graphic format is used to preserve visual appearance across different platforms, then document compatibility and visual consistency are improved, but logical structure information is lost making content reuse difficult
Solution Approach 1:
The patent segments text content into distinct text blocks with identified semantic roles (body text, titles, captions, etc.). By dividing the document content into structured segments with defined characteristics, the system recovers logical structure information that was lost in the flat vector graphic format, enabling content reuse while maintaining visual consistency.
Solution Approach 2:
The patent introduces an intermediary processing layer that analyzes vector graphic documents and generates structured output with semantic information. This intermediary system acts as a bridge between the visual representation and logical structure, extracting and organizing content meaning without altering the original visual format.
2Measurement precision
If text is divided into text blocks for structural analysis, then logical structure recognition is improved, but text may be incorrectly split into too short blocks or misclassified
Solution Approach 1:
The patent employs multiple parameters for text block classification including font type, font size, position coordinates, and semantic role characteristics. By analyzing combinations of these parameters rather than single criteria, the system achieves more accurate classification while avoiding incorrect splitting or misclassification of text blocks.
Solution Approach 2:
The system incorporates feedback mechanisms where classification results are continuously refined. Text blocks are pre-classified, then further processed and validated against expected structural patterns. Incorrect classifications are identified and corrected through iterative refinement, improving both structure recognition and classification accuracy.
Data Source
AI summary
A computer implemented method configured to analyze contents of a page of a vector graphic format file includes dividing text content on the page into text blocks, pre-classifying each text block to be one of a raw body text block and a non-body text block, processing the raw body text blocks to form a plurality of body text blocks and combining the body text blocks into a text flow including the plurality of body text blocks. The method further includes defining a semantical role of each of the non-body text blocks, and combining the non-body text blocks among the body text blocks of the text flow in a geometrical order. Result data is provided, formatted as any one of a data structure and a data stream, the result data including text content of the page as geometrically arranged non-body text blocks and body text blocks.