Document Layout Analysis for Hierarchical Entity Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in efficiently extracting and organizing technical information from heterogeneous oil and gas documents, such as lithology reports and drilling reports, due to their varied formats and content structures.
Innovation Solution
A system and method utilizing machine learning models to detect and hierarchically organize primary and secondary entities within these documents, employing preprocessing, deep learning-based object detection, and synthetic training data to create a document layout structure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are used to detect and organize entities in heterogeneous documents, then extraction accuracy and organization structure are improved, but system complexity increases
Solution Approach 1:
The system segments the document analysis task into multiple stages: preprocessing, primary entity detection (headings, tables, figures), secondary entity detection, and hierarchical organization. Each stage uses specialized machine learning models trained on specific entity types, allowing the system to handle heterogeneous documents through modular processing stages rather than a monolithic complex system
Solution Approach 2:
The patent introduces a hierarchical dimension to entity organization by detecting primary entities at one level and secondary entities at another level, creating a multi-level document layout structure. This dimensional approach allows the system to manage complexity by organizing entities in nested hierarchies rather than attempting to process all entities simultaneously at a single level
2Measurement precision
If multiple machine learning models are employed for different entity types, then detection accuracy improves, but processing time increases
Solution Approach 1:
The system performs preliminary actions by detecting primary entities (headings, tables, figures) before detecting secondary entities. The primary entity detection creates a structural framework that guides subsequent secondary entity detection, reducing the search space and computational requirements for the second stage while maintaining high accuracy through specialized models for each entity type
Solution Approach 2:
The patent implements dynamic processing where the system adapts its processing strategy based on document content. The machine learning models dynamically adjust their detection parameters and thresholds based on the specific document type and detected entity characteristics, optimizing processing speed for each document while maintaining accuracy through adaptive rather than fixed processing parameters
3Stability of the object's composition
If hierarchical organization structure is implemented, then information organization quality improves, but data processing complexity increases
Solution Approach 1:
The system implements nested organization where secondary entities are contained within and associated with primary entities in a hierarchical structure. This nesting approach allows the system to manage complexity by organizing data in nested containers rather than flat structures, with each level of hierarchy containing only the relevant entities for that level, thereby improving organization quality while managing processing complexity through structured data containment
Data Source
AI summary
A method involves detecting primary entities in a document, involving determining that a subset of the primary entities are associated with a first primary entity type, and determining a second primary entity type of one of the primary entities. The method further involves processing the primary entity of the second primary entity type to determine a secondary entity type of the primary entity. The secondary entity type is a subcategory of the second primary entity type. The method also involves hierarchically organizing the primary entities into a document layout structure that includes a top level and a child level. The top level is established by the first subset of primary entities based on the first primary entity type identifying the first subset as headings, and the child level is established by the primary entity based on the second primary entity type, the child level identifying the secondary entity type.