Bounding Box Schema for Accurate Key-Value Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data extraction methods from electronic documents, such as rule-based approaches and AI-driven techniques, face challenges in accurately and reliably extracting key-value pairs, leading to incomplete and inaccurate data extraction, potential data loss, or corruption.
Innovation Solution
A method and system that utilize two-dimensional geometrical shapes, specifically bounding boxes, to locate and extract key-value pairs from electronic documents. This approach defines bounding boxes around keys and values, employs normalization operations to handle rotated or distorted documents, and creates a searchable index for enhanced search and retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If rule-based approaches with string matches and OCR are used for data extraction, then the extraction process is simple to implement, but the extraction accuracy and reliability are insufficient
Solution Approach 1:
The patent introduces a schema file as an intermediary that defines the expected structure and location of key-value pairs in documents. This schema acts as a mediator between the simple extraction process and the complex document structures, enabling accurate extraction without complex AI models by using the schema to guide the extraction process
Solution Approach 2:
The schema file is created in advance by annotating sample documents with bounding boxes around key-value pairs. This preliminary action establishes the extraction rules before processing actual documents, allowing the system to accurately extract data from new documents without requiring complex real-time analysis
2Measurement precision
If AI-driven techniques with machine learning are used to understand document layout, then extraction accuracy may improve, but computational cost increases significantly and results become inconsistent
Solution Approach 1:
The patent replaces expensive AI models with a simple schema file-based approach. The schema file serves as a lightweight, disposable template that can be easily created and applied without requiring significant computational resources, eliminating the need for costly machine learning inference while maintaining consistent and reliable extraction results
3Productivity
If conventional extraction methods are used, then processing speed may be maintained, but data loss and corruption occur due to incomplete extraction
Solution Approach 1:
The patent transitions from point-based or text-based extraction to bounding box-based extraction that captures the spatial dimensions of key-value pairs. By using bounding boxes that encompass the complete visual area of each field, the system ensures no data is lost due to incomplete extraction, while the structured approach maintains processing efficiency
Data Source
AI summary
A method for processing one or more electronic documents for enhanced search includes defining a bounding box around each key and value of key-value pairs in a first schema file, tagging coordinates of a key corresponding to a first bounding box and coordinates of a value corresponding to a second bounding box in the first schema file. Furthermore, obtaining a first inference file, detecting coordinates of a key corresponding to a third bounding box, and determining coordinates of a fourth bounding box and the value of the first inference file that are determined by applying a normalization operation. Thereafter, extracting value encompassed by the fourth bounding box of the first inference file and automatically creating a searchable index of the first inference file with searchable key-value pairs. The method achieves an efficient and accurate clustering of data items with an accurate, meaningful, and formal objective function.