Document Data Extraction Using Region and Label Tensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy databases containing valuable data are often underutilized due to obsolete technology, limiting their ability to provide meaningful insights and increasing memory requirements.
Innovation Solution
A system and method that utilizes data from legacy databases to train an extraction algorithm, creating region and label tensors to extract target data from newer databases, improving processing speeds and reducing memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data from legacy databases is extracted and processed using traditional methods, then data can be preserved and accessed, but processing speed is slow and memory space requirements increase
Solution Approach 1:
The system performs preliminary actions by creating region tensors and label tensors from legacy database documents before actual data extraction is needed. Region tensors encode spatial information about data locations, while label tensors encode semantic information about data types. This pre-processing allows the trained algorithm to quickly extract target data from current databases without repeatedly analyzing entire documents, thus improving processing speed while reducing memory requirements.
Solution Approach 2:
The system creates simplified representations (copies) of the original documents in the form of tensors. Instead of storing and processing entire document images or texts, the system creates region tensors that capture only the essential spatial and semantic information needed for data extraction. This copying approach maintains the necessary information while dramatically reducing memory space requirements.
2Adaptability or versatility
If traditional data extraction methods are used on legacy databases, then all data can be accessed, but the technology becomes obsolete and processing efficiency decreases
Solution Approach 1:
The system introduces tensors as an intermediary representation between legacy database documents and the data extraction process. Region tensors and label tensors serve as mediators that translate legacy document formats into a standardized format that can be efficiently processed by machine learning algorithms. This intermediary approach maintains compatibility with legacy systems while enabling modern, efficient processing methods.
Solution Approach 2:
The system changes the parameters of data representation by converting traditional document formats into tensor representations. Region tensors encode spatial parameters (coordinates, dimensions) while label tensors encode semantic parameters (data types, categories). This parameter transformation enables the system to process legacy data using modern algorithms, improving processing efficiency while maintaining adaptability to legacy systems.
Data Source
AI summary
Systems and methods for enabling target data to be extracted from documents are disclosed herein. In an embodiment, a method of enabling target data to be extracted from documents includes accessing a database including a plurality of documents including target data, for each of multiple of the documents, creating a region tensor based on extracted text including the target data, for each of the multiple of the documents, creating a label tensor based on an area including the target data, and using the region tensor and the label tensor, training an extraction algorithm to extract the target data from additional documents.