Document Classification Using Physical and Lexical Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current document classification methods rely solely on lexical features, neglecting the valuable information provided by physical characteristics of documents, which can lead to inaccurate categorization, especially for disparate types and sizes of documents.
Innovation Solution
The integration of physical and lexical features for document classification, where physical attributes such as size, colorfulness, margin smoothness, and orientation are extracted and combined with lexical features to enhance classification accuracy using a neural network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If document classification relies solely on lexical features, then the classification system is simple and easy to implement, but classification accuracy deteriorates particularly for disparate types and sizes of documents
Solution Approach 1:
The patent merges lexical features (text content) and physical features (document characteristics) into a unified classification system. The physical feature extractor captures document properties such as size, orientation, and layout, while the lexical feature extractor processes text content. These two feature streams are then combined and fed into a classification algorithm that integrates both types of information to produce improved classification accuracy.
Solution Approach 2:
The patent adds another dimension to the classification problem by incorporating physical features alongside lexical features. Instead of classifying documents based solely on text content (one dimension), the system now operates in a multi-dimensional feature space that includes both lexical and physical characteristics, enabling better differentiation between document types and sizes.
2Adaptability or versatility
If physical characteristics of documents are incorporated into classification, then differentiation between document types and sizes improves, but the complexity of feature extraction and processing increases
Solution Approach 1:
The patent segments the feature extraction process into distinct modules: a physical feature extractor that analyzes document characteristics, and a lexical feature extractor that processes text content. This segmentation allows each module to specialize in specific tasks, making the overall system more manageable and easier to implement while maintaining the ability to handle diverse document types.
Solution Approach 2:
The classification system is designed with universality by accommodating both physical and lexical features within a single integrated framework. The system can handle various document types and sizes uniformly by applying the same dual-feature extraction and classification approach, making it adaptable to diverse document classification needs without requiring separate specialized systems.
Data Source
AI summary
An automatic document classification system is described that uses lexical and physical features to assign a class ciεC{c1, c2, . . . , ci} to a document d. The primary lexical features are the result of a feature selection method known as Orthogonal Centroid Feature Selection (OCFS). Additional information may be gathered on character type frequencies (digits, letters, and symbols) within d. Physical information is assembled through image analysis to yield physical attributes such as document dimensionality, text alignment, and color distribution. The resulting lexical and physical information is combined into an input vector X and is used to train a supervised neural network to perform the classification.