Machine Vision Object Recognition Model Refinement from Multiple Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine vision systems require manual and time-consuming refinement of object recognition models, leading to suboptimal performance and increased execution times due to the use of single-image or CAD-based parameter estimation, which does not account for variations in object instances.
Innovation Solution
A method for refining object recognition models by using multiple images to identify stable model points and optimizing level-specific parameters, including combining and decision functions to enhance model accuracy and robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual refinement of object recognition models is performed, then model accuracy can be improved, but time consumption and complexity increase significantly
Solution Approach 1:
The system automatically refines the object recognition model by selecting stable model points and optimizing parameters without requiring manual user intervention. The algorithm independently processes multiple images, identifies stable features, and generates an optimized model, making the system self-sufficient in the refinement process.
Solution Approach 2:
The system performs preliminary actions by pre-processing multiple images to identify stable model points before final model generation. By anticipating which points will be stable across transformations, the system prepares the model in advance, reducing the need for iterative manual refinement.
2Loss of time
If single-image or CAD-based parameter estimation is used, then setup time is reduced, but model robustness and accuracy deteriorate due to not accounting for object variations
Solution Approach 1:
The system merges information from multiple images to create a more robust object recognition model. By combining data from several images showing the same object under different conditions, the system identifies stable model points that are consistent across variations, thereby improving model reliability while maintaining efficient setup time.
Solution Approach 2:
The system changes parameters by selecting only stable model points that consistently appear across multiple images. This parameter selection process filters out unstable or variable features, creating a more robust model that generalizes better to different object instances while requiring minimal setup time.
3Measurement precision
If manual model point selection is performed, then model precision can be improved, but device complexity and user interaction requirements increase
Solution Approach 1:
The system replaces the mechanical interaction of manual model point selection with an automated computational process. The algorithm automatically identifies stable model points by analyzing multiple images and applying transformation logic, substituting human manual operations with an automated system that achieves the same or better precision without increasing user-facing complexity.
Solution Approach 2:
The system performs self-service by automatically selecting and optimizing model points without requiring user intervention. The algorithm independently evaluates multiple images, identifies stable features, and generates the final model, eliminating the need for complex user interfaces or manual configuration steps.
4Measurement precision
If exhaustive search at highest pyramid level is performed, then detection accuracy is improved, but execution time increases due to processing all levels
Solution Approach 1:
The system performs preliminary action by pre-identifying stable model points from multiple images before the actual detection process. This pre-processing step creates an optimized model that requires fewer iterations during exhaustive search, thereby maintaining detection accuracy while reducing overall execution time.
Solution Approach 2:
The system changes parameters by using stable model points selected from multiple images, which reduces the search space and number of candidates requiring exhaustive evaluation. This parameter optimization maintains detection accuracy by focusing on stable features while improving execution speed by reducing the number of computations required.
Data Source
AI summary
The present disclosure provides methods that estimate or improve various parameters of an object recognition model to improve runtime, accuracy and robustness while minimizing the required user interaction to optimize these values. In one aspect of the invention, methods are defined to generate object recognition models with refined contours. A further method is defined by the invention to estimate level-specific parameters for the object recognition algorithms.