Robust Image Identification via Feature Relationship Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image protection systems are vulnerable to distortions such as scanning, scaling, rotating, cropping, and color changes, making it difficult to identify and track the usage of images, especially when they are printed and digitized again or when they are cropped and modified.
Innovation Solution
An image comparison apparatus and method that extracts principal features from query images, establishes relationships between these features, and compares them with a catalog of images to find candidate matches, using techniques like eigenvalue normalization, multi-dimensional vector storage, and Hamming distance calculations to account for distortions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional image protection systems are used, then direct copying protection is provided, but the protection fails when images are cropped, rotated, scaled, or printed and scanned
Solution Approach 1:
The image is divided into multiple feature regions or blocks, and features are extracted from each block independently. This segmentation allows the system to maintain matching capability even when parts of the image are cropped or transformed, as long as sufficient feature blocks remain intact.
Solution Approach 2:
The system extracts and compares invariant parameters such as feature point coordinates, local gradient orientations, and scale-space characteristics that remain consistent under rotation, scaling, and illumination changes. These parameter transformations enable robust matching across different image versions.
2Adaptability or versatility
If image comparison systems are provided to match different versions, then some distortion resistance is achieved, but the systems remain vulnerable to cropping, flipping, rotating, and color balance changes
Solution Approach 1:
The system performs preliminary normalization of feature descriptors and pre-computes invariant characteristics before comparison. By preparing transformed feature representations in advance and establishing comparison protocols for various transformations, the system achieves both robustness and precision simultaneously.
3Measurement precision
If feature extraction and comparison is performed on all images, then comprehensive matching is achieved, but computational complexity and processing time increase
Solution Approach 1:
The image processing is segmented into distinct stages: feature detection, descriptor extraction, and comparison. Each stage processes only relevant data, reducing overall computational load while maintaining comprehensive matching capability.
Solution Approach 2:
The system extracts a selective subset of salient features rather than processing all image data. By focusing on the most discriminative feature points and their local descriptors, the system achieves comprehensive matching with reduced computational complexity.
Data Source
AI summary
Apparatus for matching a query image against a catalog of images, comprises: a feature extraction unit operative for extracting principle features from said query image; a relationship unit operative for establishing relationships between a given principle feature and other features in the image, and adding said relationships as relationship information alongside said principle features; and a first comparison unit operative for comparing principle features and associated relationship information of said query image with principle features and associated relationship information of images of said catalog to find candidate matches.