Composite Image Descriptors for Visual Target Differentiation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image descriptors are inadequate for accurately differentiating between a large number of similar targets, leading to unacceptably high false positive matches, particularly when searching for thousands or millions of targets in visual data management and augmented reality applications.
Innovation Solution
The development of composite image descriptors, which are formed by strategically combining pairs or larger groupings of image descriptors to generate more detailed and unique descriptors, using a system that selectively filters and combines primary and secondary image descriptors based on geometric proximity and dissimilarity, thereby reducing false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image descriptors are used to manage visual data, then the system is simple and easy to operate, but the ability to differentiate between similar targets is insufficient leading to high false positive matches
Solution Approach 1:
The patent combines multiple individual image descriptors into a composite image descriptor that integrates features from multiple detected targets. This merging approach increases the uniqueness and differentiation capability of the descriptor, allowing the system to accurately distinguish between similar targets while maintaining operational simplicity through a unified descriptor structure.
2Reliability
If composite image descriptors combining multiple descriptors are used, then differentiation accuracy improves, but computational complexity and processing time increase
Solution Approach 1:
The system performs preliminary actions by detecting multiple targets and generating individual descriptors for each target before combining them into a composite descriptor. This preliminary processing organizes the data in advance, enabling efficient composite descriptor generation that maintains high match accuracy while optimizing processing speed through structured pre-computation.
3Reliability
If multiple image descriptors are combined to form composite descriptors, then false positives are reduced, but the computational resources required increase
Solution Approach 1:
The patent applies partial action by selectively combining descriptors from detected targets based on relevance and confidence criteria. Rather than combining all possible descriptors, the system strategically selects and combines only the most relevant descriptors, reducing computational energy consumption while maintaining the reliability benefit of reduced false positives through targeted descriptor composition.
Data Source
AI summary
An illustrative image descriptor generation system determines a subset of image descriptors from a plurality of image descriptors that each correspond to a different feature point included within an image. The subset of image descriptors is determined based on geometric proximity, within the image, of respective feature points of the subset of image descriptors to a feature point of a primary image descriptor. The image descriptor generation system then selects a secondary image descriptor from the subset of image descriptors and combines the primary image descriptor and the secondary image descriptor to form a composite image descriptor. Corresponding methods and systems are also disclosed.