Image Processing Device for Cross-Modality Similarity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to accurately determine similarity between images of different modalities, such as visible and far-infrared images, due to disparities in pixel value ranges, leading to ineffective similarity indicators like absolute sum of pixel values.
Innovation Solution
An image processing device and method that deform images in multiple ways, calculate differences using various similarity evaluation methods, normalize these differences, and integrate them to assess similarity accurately across different image types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple similarity evaluation methods are used to improve accuracy across different image types, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent segments the similarity evaluation process into multiple independent evaluation methods (e.g., absolute difference, correlation coefficient, mutual information), each handling specific image type combinations. This allows the system to select appropriate evaluation methods based on image modality pairs, improving accuracy without requiring all methods to run simultaneously, thus managing complexity.
Solution Approach 2:
The patent creates a universal similarity evaluation framework that can handle multiple image type combinations (visible-visible, visible-far-infrared, far-infrared-far-infrared) using a set of standardized evaluation methods. This multi-functional approach allows the same system to adapt to different image pairs without requiring separate processing pipelines, improving versatility while maintaining manageable complexity through standardized interfaces.
2Productivity
If absolute sum of pixel values is used for similarity evaluation, then processing speed improves, but measurement precision deteriorates when comparing images of different modalities
Solution Approach 1:
The patent changes the evaluation parameter from absolute pixel value sum to normalized metrics such as correlation coefficient and mutual information. These parameters are scale-invariant and can handle different value ranges of visible and far-infrared images. The system calculates these normalized parameters instead of raw pixel differences, improving accuracy for cross-modality comparison while maintaining computational efficiency through optimized algorithms.
Solution Approach 2:
The patent introduces normalized similarity metrics as intermediary measures between raw pixel values and final similarity assessment. These intermediary metrics (correlation coefficient, mutual information) bridge the gap between different image modalities by transforming raw pixel data into comparable standardized scores, enabling accurate cross-modality comparison without directly comparing incompatible pixel value ranges.
Data Source
AI summary
In order to ensure that similarity is detected with high accuracy, regardless of the types of images constituting a group of images to be evaluated for similarity, the image processing device includes difference calculation means 11 for deforming one of two or more images constituting an image group in one or more deforming ways, and calculating a degree of difference between the deformed image and the other image or images in the image group for each pixel using multiple ways for similarity evaluation, normalization means 12 for normalizing each degree of difference by each of the multiple ways for similarity evaluation, and difference integration means 15 for integrating normalized degrees of difference.