Document Subimage Feature Matching and Distortion Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hand-held document imaging devices produce digital images with noise, optical blur, non-standard orientation, and irregular backgrounds, leading to defects that degrade the performance of optical-character-recognition methods and systems, making it difficult to accurately extract text from these images.
Innovation Solution
The methods and subsystems identify and characterize document subimages by extracting features from images, matching them with pre-modeled features, and correcting distortions to facilitate data extraction from the subimages, using techniques like RANSAC-like model selection and SIFT feature detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If hand-held document imaging devices are used, then cost and user accessibility are improved, but image quality deteriorates due to noise, optical blur, and non-standard orientation
Solution Approach 1:
The system performs preliminary actions by detecting document boundaries and extracting document subimages before the actual OCR processing. It also performs preliminary distortion correction and noise reduction on the extracted subimages to prepare them for subsequent optical-character-recognition operations, thereby improving image quality without requiring expensive imaging hardware.
Solution Approach 2:
The patent introduces an intermediary processing pipeline between the hand-held imaging device and the OCR system. This intermediary includes steps for detecting document regions, extracting subimages, correcting distortions, and reducing noise, which mediates the poor image quality from hand-held devices to produce suitable input for accurate OCR recognition.
2Extent of automation
If computational optical-character-recognition methods are applied to hand-held images, then electronic document generation is achieved, but recognition accuracy deteriorates due to image defects
Solution Approach 1:
The system performs preliminary processing actions including distortion correction, noise reduction, and contrast enhancement on the hand-held images before feeding them to the OCR system. This preliminary action prepares the images to maintain high recognition accuracy despite the automated generation process.
Solution Approach 2:
The patent employs feedback mechanisms where the system detects document boundaries and validates extracted subimages against expected document characteristics. If quality issues are detected, the system can re-process or adjust parameters to ensure reliable recognition accuracy throughout the automated document generation pipeline.
3Measurement precision
If document subimages are extracted and corrected, then optical-character-recognition accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent segments the complex processing task into distinct modular steps: document boundary detection, document subimage extraction, distortion correction, noise reduction, and OCR processing. This segmentation allows each step to be optimized independently and simplifies the overall system architecture while maintaining high recognition accuracy.
Solution Approach 2:
The system applies local quality enhancement by extracting only the relevant document portions (subimages) and applying distortion correction and noise reduction specifically to these regions rather than the entire image. This localized approach improves OCR accuracy while minimizing the computational complexity associated with processing the whole image.
Data Source
AI summary
The present document is directed to methods and subsystems that identify and characterize document-containing subimages in a document-containing image. In one implementation, each type of document is modeled as a set of features that are extracted from a set of images known to contain the document. To locate and characterize a document subimage in an image, the currently described methods and subsystems extract features from the image and then match model features of each model in a set of models to the extracted features to select the model that best corresponds to the extracted features. Additional information contained in the selected model is then used to identify the location of the subimage corresponding to the document and to process the document subimage to correct for a variety of distortions and deficiencies in order to facilitate subsequent data extraction from the corrected document subimage.