Document Layout Grouping Using Unsupervised OCR Feature Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for grouping documents based on layout detection are inefficient and require significant manual effort, especially for documents without predefined formats, such as financial documents, due to the complexity of identifying unique layouts for similar data extraction tasks.
Innovation Solution
A method and system utilizing an unsupervised machine learning model to determine document layouts by combining text and image features, using OCR, encoder-decoder deep learning, and unsupervised machine learning to automatically group documents based on their layout similarities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual document grouping is performed, then documents can be categorized, but the process becomes onerous and time-consuming
Solution Approach 1:
The patent replaces manual mechanical document grouping with an automated computer vision system that uses image processing algorithms to detect document layouts, extract text regions, and categorize documents automatically, eliminating the need for manual intervention while maintaining high accuracy
Solution Approach 2:
The patent introduces an intermediary AI model that acts as a bridge between raw document images and final grouping results, using trained machine learning models to perform layout detection and classification, thereby reducing both time and manual effort
2Adaptability or versatility
If separate document layouts are used for different data extraction purposes, then specific extraction tasks can be performed, but it becomes cumbersome to identify unique layouts
Solution Approach 1:
The patent creates a universal document layout detection system that can handle multiple data extraction purposes through a single unified model, capable of identifying various document types (invoices, receipts, forms, etc.) and their layouts using the same core technology, thereby reducing complexity while maintaining versatility
Solution Approach 2:
The patent uses parameter changes in the form of adjustable classification categories and configurable extraction fields, allowing the same layout detection system to adapt to different data extraction needs by changing parameters rather than requiring separate layouts for each purpose
3Extent of automation
If AI/ML techniques are used for document grouping, then automation is achieved, but efficiency remains insufficient
Solution Approach 1:
The patent segments the document processing task into distinct stages: image preprocessing, text region detection, layout analysis, and final grouping, with each stage handled by optimized sub-models that work together efficiently, improving overall productivity while maintaining high automation
Solution Approach 2:
The patent performs preliminary actions by pre-processing document images (normalization, noise removal, enhancement) before main analysis, and by pre-training models on diverse document datasets, enabling faster and more efficient processing during actual document grouping operations
Data Source
AI summary
A method and system of grouping a plurality of documents is disclosed. The method includes determining a plurality of text features of each of the plurality of documents. A plurality of image features of each of the plurality of documents are determined. A layout features set of each of the plurality of documents are determined based on the plurality of text features and the plurality of image features. Each of the plurality of documents in one group from a set of groups is grouped based on determination of a document layout of each of the plurality of documents.