Drawing Title Field Recognition for Similar Drawing Search

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing drawing management systems struggle to accurately specify attribute values in diverse title fields of drawing data due to the lack of a unified standard, leading to difficulties in accurately searching for similar drawing data.

Innovation Solution

A drawing search system that specifies the type of title field, arranges boxes within the field, performs character recognition, and corrects the recognition results using predefined rules and dictionaries to accurately identify and associate attribute values, enabling precise searching of similar drawing data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If character recognition processing is performed on title fields without specifying type and arrangement, then processing speed is maintained, but attribute value specification accuracy deteriorates

Engineering Contradiction:

Improveattribute value specification accuracyVSAvoidtitle field analysis complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The title field is segmented into multiple boxes with specific arrangements (e.g., first box for item name, second box for attribute value). This segmentation allows the system to apply different processing rules to different regions, improving attribute value specification accuracy while keeping the overall process manageable through structured division.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary classification of title field types and determination of box arrangements before executing character recognition. This preliminary action establishes a framework that guides subsequent recognition processing, ensuring that attribute values are extracted from the correct positions with appropriate context, thereby improving accuracy without significantly increasing overall complexity.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If title fields are processed without type specification, then processing time is reduced, but search accuracy for similar drawing data deteriorates

Engineering Contradiction:

Improvesearch accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

Different processing approaches are applied to different types of title fields based on their specific characteristics. The system identifies title field types and applies localized processing rules appropriate for each type, which improves search accuracy by capturing domain-specific patterns while avoiding the need for a single overly complex processing pipeline that would increase processing time.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system changes processing parameters based on title field type and box arrangement. By dynamically adjusting recognition parameters, search criteria, and analysis depth according to the specific structure identified, the system achieves high search accuracy for similar drawing data while optimizing processing time through adaptive parameter selection rather than uniform high-cost processing.

Inventive Principle:

Principle #35Parameter changes

3Loss of information

If comprehensive character recognition is performed on all title field content, then information completeness is improved, but difficulty in identifying specific attribute values increases

Engineering Contradiction:

Improveinformation completenessVSAvoidattribute value identification difficulty

Core Design Contradiction:

Loss of informationVSDifficulty of detecting and measuring

Solution Approach 1:

The title field content is segmented into distinct boxes with defined functions (item name box, attribute value box, etc.). This segmentation preserves all information while making it easily locatable and identifiable, as each box has a predetermined role. The system can comprehensively process all content while the structured segmentation prevents confusion about which elements are attribute values.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system extracts attribute values from their specific box positions after recognizing the title field structure. By taking out attribute values from the broader title field context and placing them in a standardized extraction framework, the system maintains information completeness while significantly reducing the difficulty of identifying and processing specific attribute values for search purposes.

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS20260064765A1Drawing search device, drawing database construction device, drawing search system, drawing search method, and recording medium

Publication Date: 2026.03.05 CADDI INC

AI summary

A drawing search server specifies an arrangement of plural boxes existing in a title field title field existing in target drawing data. The drawing search server executes character recognition processing on each character string existing at each position in the title field existing in the target drawing data. Based on the character recognition result and the arrangement of the plural boxes, the drawing search server specifies a target attribute value representing an attribute value of the target drawing data and specifies a target attribute which is an attribute associated in advance with the box. The drawing search server searches a drawing database for drawing data having an attribute value similar to the attribute value of the target drawing data based on a combination of the target attribute value and the target attribute, and outputs a search result.