Genetic Programming Image Filter Fitness via Parameter Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing filter production methods using genetic programming often fail to select effective filters due to variations in edge line widths and teaching errors, leading to suboptimal fitness calculations and potential exclusion of effective processes.
Innovation Solution
An image processing filter producing apparatus that genetically evolves filters by producing new filters, acquiring parameters that identify shapes in input and output images, and calculating fitness through comparing these parameters to ensure robustness against environmental changes and improve selection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional fitness calculation compares images or pixels before and after filtering, then the process is simple to implement, but variation of edge line widths and teaching errors decrease fitness accuracy causing effective filters to be excluded
Solution Approach 1:
The patent changes the parameter basis for fitness calculation from direct image/pixel comparison to comparison of extracted parameters (such as edge line positions, shapes, or features). This transforms the fitness evaluation from comparing raw image data to comparing extracted parameter data, thereby improving accuracy while managing complexity through parameter extraction and comparison
2Adaptability or versatility
If genetic programming evolves image processing filters through crossover and mutation, then new filters are generated to adapt to environmental changes, but the process requires repeated generations and extensive computation time
Solution Approach 1:
The patent applies preliminary action by pre-defining the structure and parameters of image processing filters before the genetic programming evolution begins. By establishing the filter framework in advance and using parameter-based fitness evaluation, the system reduces the computational burden during evolution, allowing faster convergence while maintaining adaptability to environmental changes
3Measurement precision
If a worker manually selects image processing filters by comparing output images, then selection accuracy depends on human judgment, but the process is time-consuming and cannot easily adapt to frequent environmental changes
Solution Approach 1:
The patent replaces the mechanical human judgment system with an automated parameter-based evaluation system. Instead of workers manually comparing output images, the system automatically extracts parameters from images and compares them to evaluate filter fitness, thereby maintaining high selection accuracy while dramatically improving productivity and enabling rapid adaptation to environmental changes
Data Source
AI summary
In order to produce an image processing filter by utilizing genetic programming, a taught parameter acquiring unit acquires a taught parameter indicating a feature shape in an input image before processing. A data processing unit creates an output image by processing the input image with an image processing filter, and subsequently a feature extracting unit extracts a detected parameter indicating a feature shape in the output image. An automatic configuring unit evaluates the image processing filter by calculating cosine similarity between the taught parameter and the detected parameter.