Multi-source Image Correspondence via Heterogeneous Model Fitting

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing multi-source image correspondence methods face challenges due to sensitivity to linear intensity variations and low local repeatability of features, limiting their application scope and accuracy.

Innovation Solution

A multi-source image correspondence method based on heterogeneous model fitting, which constructs a multi-orientation phase consistency model using log-Gabor filters and variable-size bins, alleviates nonlinear radiation distortion and abnormal matching relationships, improving feature detection accuracy and robustness.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If SIFT or improved SIFT variants are used for feature extraction, then robustness to scale and rotation is improved, but sensitivity to linear intensity variations increases and local repeatability decreases

Engineering Contradiction:

Improverobustness to scale and rotationVSAvoidlocal repeatability

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent changes the fundamental parameters of feature extraction by using phase consistency in the frequency domain instead of intensity-based methods. The phase spectrum is invariant to intensity variations, and the patent computes phase consistency across multiple orientations and scales, fundamentally changing how features are detected and described, thereby achieving both robustness and repeatability

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mechanical/intensity-based feature extraction approach (SIFT) with a frequency-domain phase-based approach. By transforming to the frequency domain and using phase information, the system substitutes the intensity-based mechanism with a phase-based mechanism that is inherently more robust to intensity variations while maintaining geometric invariance

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If traditional feature extraction methods are used, then processing speed is maintained, but accuracy of multi-source image correspondence decreases due to low local repeatability

Engineering Contradiction:

Improveprocessing speedVSAvoidcorrespondence accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the feature extraction process into distinct frequency-domain operations: computing the phase spectrum, calculating phase consistency across orientations, and generating descriptors. This segmentation allows efficient parallel computation while maintaining high accuracy, resolving the contradiction between speed and precision

Inventive Principle:

Principle #1Segmentation

3Device complexity

If single basic transformation model is used, then device complexity is reduced, but performance of image correspondence is limited

Engineering Contradiction:

Improvemodel complexityVSAvoidimage correspondence performance

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent creates a universal heterogeneous model that can handle multiple types of geometric transformations (rigid, affine, perspective) within a single framework. The model is designed to be multi-functional, adapting to different transformation types based on the data, thereby improving performance without proportionally increasing complexity

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent combines multiple basic transformation models into a composite heterogeneous model. By fusing different transformation models (rigid, affine, perspective) into a unified framework, the system achieves the benefits of multiple models while managing complexity through a structured composite approach

Inventive Principle:

Principle #40Composite materials

Data Source

PatentUS12131517B1Multi-source image correspondence method and system based on heterogeneous model fitting

Publication Date: 2024.10.29 JINAN UNIVERSITY

AI summary

A multi-source image correspondence method and system based on heterogeneous model fitting is provided, the method includes the following steps: constructing a multi-orientation phase consistency model, fusing phase consistency, image amplitude, and orientation detection feature points, constructing logarithmic polar coordinate descriptors with variable-size bins using sub-region grids and orientation histograms, effectively estimating model parameters through heterogeneous model fitting, accumulating matching pairs from different heterogeneous models that meet a preset joint position offset transformation error, outputting a final matching pair, and completing multi-source image correspondence. The present disclosure alleviate the influence of nonlinear radiation distortion by constructing the multi-orientation phase consistency model, constructing logarithmic polar coordinate descriptors with variable-size bins by sub-region grids and orientation histograms, removing an abnormal matching relationship in multi-source images with the heterogeneous model fitting method, thereby improving the accuracy and robustness of feature detection and improving multi-source image correspondence performance.