Document Registration via 2D Projection Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document registration algorithms are inadequate for registering historical documents that are damaged or deteriorated due to age, poor preservation, and natural or man-made disasters, as they rely on locating corners or key points, which are unreliable for severely damaged documents.
Innovation Solution
Adaptation of dynamic time warping (DTW) algorithms from speech recognition to register damaged documents by projecting 2D images onto axes, creating 1D signals, and using correlation functions to align and weight projections, making the method robust to variations in scale, translation, and damage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If corner-based or key-point-based registration algorithms are used, then registration accuracy is improved for intact documents, but reliability deteriorates for damaged documents
Solution Approach 1:
The patent segments the 2D document image into multiple 1D projections along different axes (horizontal, vertical, and diagonal). This segmentation transforms the problem from locating specific key points in 2D space to analyzing distributed projection patterns, making the registration process robust to localized damage while maintaining accuracy through cumulative evidence from multiple projection directions.
2Loss of information
If 2D image processing methods are used, then detailed document features are preserved, but computational complexity increases
Solution Approach 1:
The patent reduces computational complexity by transforming the 2D image processing problem into multiple 1D projection problems. By projecting the document image and form images onto different axes, the method simplifies the computational burden while preserving essential structural information through the distribution of features across multiple projection dimensions.
3Productivity
If traditional registration methods are used, then processing speed is maintained, but adaptability to damaged documents deteriorates
Solution Approach 1:
The patent performs preliminary projections of both the damaged document image and reference form images onto multiple axes before attempting registration. This preliminary action creates a set of 1D projection patterns that are more tolerant to damage, allowing the correlation-based registration to proceed effectively even when significant portions of the original 2D image are damaged or missing.
Data Source
AI summary
Systems and methods for registering documents. A two-dimensional document image and one or more form images may be obtained. The document image may be projected onto a horizontal axis and a vertical axis to create a horizontal document projection and a vertical document projection. The form images may be projected onto the horizontal axis and the vertical axis to create a horizontal form projection and a vertical form projection. The horizontal document projection may be correlated with the horizontal form projection of the form images and the vertical document projection may be correlated with the vertical form projection of the form images. Correlation scores may be calculated based on the correlations.