Image Text Extraction Using Lexicon-Guided Word Splitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical character recognition (OCR) technologies struggle with inaccurate text extraction from low-quality images, leading to errors in identifying and correcting textual information, particularly in documents with distorted or degraded quality.
Innovation Solution
A computer-implemented method that utilizes domain and language-specific knowledge databases, including domain lexicon and language dictionary databases, to compare and correct extracted text, employing image pre-processing techniques like adaptive thresholding, image enhancement, and de-skewing, followed by splitting concatenated words to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional OCR engine is used for text extraction, then the process is simple and fast, but the text extraction accuracy is low especially for low-quality images
Solution Approach 1:
The system segments the text extraction process into multiple stages: initial OCR extraction, word validation against domain lexicon and language dictionary, error identification, and corrective processing. This segmentation allows each stage to focus on specific aspects, improving overall accuracy while managing complexity through modular design.
Solution Approach 2:
The system performs preliminary validation of extracted words against domain-specific lexicons and language dictionaries before final output. This preliminary action identifies potential errors early in the process, allowing for targeted correction rather than attempting to improve overall OCR accuracy through more complex initial processing.
2Measurement precision
If image quality is improved through pre-processing, then text extraction accuracy increases, but processing time and computational resources increase
Solution Approach 1:
Instead of applying comprehensive image pre-processing to all images, the system applies validation and correction only to extracted text that fails to match domain lexicons or language dictionaries. This partial action approach processes only the necessary portions, reducing overall processing time while maintaining accuracy for problematic cases.
Solution Approach 2:
The system uses domain-specific knowledge databases and language dictionaries to self-correct extraction errors without requiring manual intervention or extensive pre-processing. The validation mechanism automatically identifies and corrects errors by comparing against known correct terms, reducing the need for time-consuming image quality improvements.
3Measurement precision
If domain-specific knowledge databases are integrated, then text extraction accuracy improves through better word matching, but system complexity and memory requirements increase
Solution Approach 1:
The system implements local quality by using domain-specific lexicons tailored to particular industries or contexts rather than a single comprehensive database. This allows the system to load only the relevant domain knowledge needed for each specific task, reducing memory requirements while maintaining high matching accuracy for the intended application domain.
Data Source
AI summary
Embodiments of the present disclosure provide systems and methods for performing text extraction from an image including textual data. The method performed by a processor includes extracting machine-readable textual data from the image. The machine-readable textual data includes one or more words. The method includes comparing each of the one or more words with a dataset including a domain lexicon database and a language dictionary database to determine a first set of words and a second set of words. The first set of words is words successfully matching with words available in the dataset, and the second set of words is words with no successful match with words available in the dataset. Further, the method includes splitting at least one word of the second set of words into two or more words to determine a third set of words and generating a textual output associated with the image.