Document Attribute Recognition for Database-Light Content Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical character recognition (OCR) technologies are resource-intensive and economically infeasible for smaller clients due to the need for maintaining large databases of historical documents, which is costly and often prohibited by confidentiality concerns.
Innovation Solution
Implementing model-based attribute detection using trained neural networks to identify document attributes without relying on large databases, allowing efficient extraction of relevant information from documents, even with limited sample data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large databases of historical documents are used for OCR, then recognition accuracy is improved, but cost and resource requirements increase significantly
Solution Approach 1:
The patent extracts only the essential attribute information (text content, structure, metadata) from historical documents and stores them in compact attribute databases rather than storing complete document copies. This extraction approach maintains recognition accuracy by preserving key identifying features while dramatically reducing storage requirements and resource consumption.
Solution Approach 2:
Instead of storing complete document images and full-text content in databases, the patent inverts the approach by storing only extracted attribute data (metadata, structural information, key text fields). The system then uses these attributes for recognition and matching, achieving accurate document identification without the resource burden of storing complete document copies.
2Reliability
If database-based OCR is implemented, then document recognition capability is improved, but system complexity and maintenance burden increase
Solution Approach 1:
The patent segments the document processing system into distinct modules: attribute extraction module, attribute database module, and recognition module. Each module performs a specific function - extracting relevant attributes from documents, storing them in structured databases, and using them for recognition. This segmentation reduces system complexity by making each component independent and manageable while maintaining overall recognition capability.
Solution Approach 2:
The patent performs preliminary attribute extraction and organization during document ingestion, preparing structured attribute data in advance before recognition is needed. This preliminary action creates ready-to-use attribute databases that speed up the recognition process and reduce the computational complexity during actual document matching operations.
3Measurement precision
If complete document databases are maintained for each client, then client-specific recognition accuracy is improved, but confidentiality and security risks increase
Solution Approach 1:
The patent extracts only non-sensitive attribute information (metadata, structural attributes, anonymized text fields) from client documents while leaving out confidential content. These extracted attributes are stored in databases that can be shared across clients for improved recognition accuracy, while the actual confidential document content remains private and is not stored in the shared databases, thus reducing confidentiality risks.
Data Source
AI summary
Aspects and implementations provide for techniques of fast and efficient recognition of texts in electronic documents. The disclosed techniques include, for example, processing, a document to obtain a first (second, etc.) set of hypotheses each associating the document with a respective value of a first (second, etc.) document attribute, form a combined hypotheses each including hypotheses of the first set and the second set. The techniques further include identifying a preferred combined hypothesis associating a first value with the first document attribute and a second value with the second document attribute, and extracting, using the first value and the second value, information content of the document.