Object Matching Model Using Sub-Feature Contribution Indicators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object matching algorithms often produce incorrect matches due to interference and noise, making it difficult to distinguish correct from incorrect object locations in images.
Innovation Solution
A method to generate a second object model by reducing the influence of sub-features that contribute most to incorrect matches, achieved by obtaining sub-models and contribution indicators based on matching with model optimization images, allowing for a more accurate and efficient object matching process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all object features are included in the object model for comprehensive matching, then matching coverage is improved, but incorrect matches increase due to interference and noise
Solution Approach 1:
The patent segments the object model into multiple sub-models, each representing different parts or aspects of the object features. This segmentation allows selective weighting and evaluation of different feature subsets, enabling the system to identify and reduce the influence of sub-features that contribute to incorrect matches while maintaining comprehensive coverage through the combined sub-models.
Solution Approach 2:
The patent introduces contribution indicators that quantify the impact of each sub-feature on matching results. By calculating and adjusting weights based on these indicators, the system dynamically changes the parameters (weights) of different sub-features in the object model, reducing the influence of noisy or interfering features while enhancing reliable features.
2Reliability
If comprehensive object features are used to ensure accurate matching, then matching reliability is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary evaluation by matching sub-models with model optimization images before final object matching. This preliminary action identifies which sub-features contribute to incorrect matches, allowing the system to pre-calculate contribution indicators and establish optimal weights. This preparation reduces processing time during actual matching operations by avoiding unnecessary computations with low-value features.
Solution Approach 2:
The patent evaluates and weights sub-features based on their contribution to correct versus incorrect matches. By focusing computational resources on sub-features with high contribution indicators and reducing or eliminating those with low indicators, the system performs partial action on the most valuable features, achieving accurate matching with reduced processing time.
3Productivity
If simple object features are used to reduce processing complexity, then processing speed is improved, but matching precision deteriorates
Solution Approach 1:
The patent transforms simple object features into weighted contributions through the contribution indicator mechanism. By calculating the relative importance of each sub-feature based on matching results with model optimization images, the system enhances the effective precision of simple features without adding computational complexity of complex feature extraction.
Solution Approach 2:
The patent uses feedback from matching sub-models with model optimization images to evaluate and adjust the weight of each sub-feature. This feedback loop identifies which simple features contribute to correct matches and which lead to incorrect matches, allowing the system to optimize the use of simple features for both speed and precision.
Data Source
AI summary
Device(s) and method supporting generation of a second object model, based on a first object model, for object matching according to an object matching algorithm. The first object model comprising object features of an imaged reference object. It is obtained sub-models that comprise different sub-features, respectively, of said object features comprised in the first object model. It is provided contribution indicators for the sub-models, respectively. Each contribution indicator indicating contribution of the sub-feature to incorrect matches. The contribution indicators being based on matching, according to the object matching algorithm, the first object model and the sub-models with at least one model optimization image comprising predefined training features that when matched with the first object model result in at least said incorrect matches.