Table Data Comparison Using Cell Row and Table Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing character reading devices cannot determine changes in table items before and after revisions, as they only recognize characters and not structural differences in tables.
Innovation Solution
A data comparison method that acquires character strings from cells in two tables, calculates cell, row, and table similarities using similarity expressions, and identifies differences based on predetermined thresholds to determine correspondence and highlight changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If character reading devices only recognize characters, then character identification is achieved, but table structure changes and item differences cannot be determined
Solution Approach 1:
The patent segments the table comparison task into multiple levels: cell-level comparison (individual character strings), row-level comparison (sequences of cells), and table-level comparison (overall structure). This segmentation allows the system to preserve both character recognition accuracy and table structure information by analyzing differences at each hierarchical level independently.
Solution Approach 2:
The patent introduces a structural dimension to the comparison process by evaluating table correspondence based on multiple criteria: cell correspondence, row correspondence, and overall table structure. This multi-dimensional approach transforms the single-dimensional character recognition task into a comprehensive analysis that captures both content and structural changes.
2Device complexity
If character reading devices process only individual characters, then character identification is simple, but table item differences and structural changes cannot be identified
Solution Approach 1:
The patent divides the complex table comparison task into manageable segments: extracting cell character strings, comparing cells individually, grouping cells into rows, comparing rows, and finally determining table-level correspondence. This segmentation reduces processing complexity by breaking down the overall task into smaller, more manageable sub-tasks that can be executed systematically.
Solution Approach 2:
The patent performs preliminary actions by first extracting and organizing cell character strings before comparison, and by determining cell and row correspondences before performing the final table-level comparison. These preliminary steps prepare the data in advance, making the subsequent comparison process more efficient and accurate.
3Measurement precision
If table comparison requires identifying all differences, then complete change detection is achieved, but processing time and computational resources increase
Solution Approach 1:
The patent segments the comparison process to identify differences at multiple levels (cell, row, table), allowing the system to stop at the appropriate level of detail needed. This segmentation enables efficient processing by focusing computational resources on the most relevant level of comparison rather than exhaustively analyzing all possible differences.
Solution Approach 2:
The patent implements a multi-level comparison approach where the system can perform partial comparisons at cell and row levels before proceeding to full table-level analysis. This allows the system to identify obvious differences quickly at lower levels, reducing the need for exhaustive full-table comparison in many cases.
Data Source
AI summary
An information processing device acquires character strings in respective cells that are included in a first table and a second table in document data. When acquiring a character string, the device may identify the characters included in the character string by using a pre-trained model that has been trained in advance through machine learning. The device determines whether the second table corresponds to the first table based on similarities between character strings in cells that are included in the first table and character strings in cells that are included in the second table. When determining that the second table corresponds to the first table, the device identifies a difference between the character strings in the cells that are included in the first table and the character strings in the cells that are included in the second table and correspond to the cells included in the first table.