Heuristics-Based Document Content Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optical Character Recognition (OCR) systems struggle to automatically identify logical connections and relationships between textual elements in documents, especially when graphic symbols are absent or not physically close, requiring significant human supervision and effort for training machine learning models.
Innovation Solution
A heuristics-based approach that mimics human perception by using spatial clues from textual and graphic symbols to cluster elements, associate textual elements as 'keys' with attributes, and generate a structured representation for automated document processing, which can be used to train machine learning models without prior knowledge of the form's structure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional OCR systems are used to extract text from documents, then text characters can be recognized, but logical connections and relationships between textual elements cannot be automatically identified
Solution Approach 1:
The document is segmented into multiple visual elements (textual elements, graphic elements, and container elements) with distinct hierarchical levels. Each element is analyzed separately for its properties and relationships, allowing the system to identify logical connections by examining spatial and structural segmentations rather than treating the document as a uniform text stream.
Solution Approach 2:
The system transitions from one-dimensional linear text processing to two-dimensional spatial analysis by considering the positions, orientations, and hierarchical nesting of visual elements. This dimensional shift enables automatic identification of logical relationships through spatial clustering and container-element associations without requiring human supervision.
2Extent of automation
If machine learning models are trained to recognize document structures, then automated processing can be achieved, but significant human supervision and effort are required for training
Solution Approach 1:
The system performs self-service by automatically analyzing visual elements and their spatial relationships to identify document structures and logical connections. The heuristic algorithms autonomously process documents without requiring human-labeled training data, eliminating the time-consuming manual supervision typically needed for training machine learning models while achieving automated document processing.
3Reliability
If graphic symbols are used to indicate contextual relationships, then fields can be grouped logically, but documents without such symbols or with non-physical proximity require complex processing
Solution Approach 1:
The system implements a universal approach that handles both documents with explicit graphic symbols and those without. By analyzing spatial clustering, container-element relationships, and visual element properties, the same heuristic algorithms can reliably identify logical groupings regardless of whether graphic symbols are present, physical proximity exists, or explicit links are provided, thereby reducing processing complexity across diverse document types.
Data Source
AI summary
A computer-implemented method for obtaining content of a document is provided. The method includes: receiving data in an unknown format obtained by an OCR application from the document, the data comprising a plurality of visual elements; for each of the plurality of visual elements, obtaining a position in the document; determining, from the plurality of visual elements, one or more graphic elements and one or more textual elements; determining a particular graphic element from the one or more graphic elements based on the position of the particular graphic element; determining, from the one or more textual elements, a key that is associated with the particular graphic element; determining, from the one or more textual elements, one or more attributes that are associated with the particular graphic element; generating an association between the key and each of the one or more attributes; and providing a structured representation of the association.