Image Examination Region Extraction via Parameter Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image examination apparatuses face challenges in accurately setting the examination region, especially for objects with complex shapes or individual differences, and when foreground and background colors are similar, leading to reduced examination accuracy due to the influence of lighting and noise.
Innovation Solution
An image examination method that uses examination region definition information as a constraint condition for an optimum solution search process, determining the position and shape of the examination region based on pixel separation and edge overlap, with adjustable balance parameters to prioritize color or edge information, enhancing extraction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If binarization or gamut extraction is used to automatically set the examination region, then the setting work is simplified and complicated shapes can be responded to, but the extraction accuracy is degraded when foreground and background colors are similar or when lighting noise is present
Solution Approach 1:
The patent transforms the examination region setting from direct pixel-based binarization/gamut extraction to parameter-based constraint optimization. By changing the approach from local pixel analysis to global parameter optimization (using parameters like examination region area ratio, position coordinates, and shape constraints), the system achieves both automatic setting and high accuracy even when foreground and background colors are similar. The parameter-based method is less sensitive to lighting noise and color variations.
Solution Approach 2:
The patent performs preliminary extraction of edge information and to-be-examined object information before determining the final examination region. By pre-processing the image to extract edges and object contours, and using these as constraints in the optimization process, the system establishes a foundation that guides the parameter optimization toward accurate regions, avoiding the need for direct color-based segmentation that fails under similar color conditions.
2Extent of automation
If the examination region is narrowed to avoid individual differences and position/orientation fluctuations, then automatic examination can be performed, but the risk of examination omission increases
Solution Approach 1:
The patent implements dynamic examination region determination by performing parameter optimization for each to-be-examined object individually. Rather than using a fixed, pre-defined region that must accommodate all variations, the system dynamically calculates the optimal examination region parameters (position, size, shape) based on each object's specific characteristics extracted from the image. This allows the examination region to adapt to individual differences and position/orientation variations while maintaining high automation and avoiding examination omissions.
Solution Approach 2:
The system enables each to-be-examined object to define its own examination region through automatic parameter optimization based on its image characteristics. The examination region is self-determined for each object rather than being imposed from outside, allowing the system to automatically adapt to individual variations without manual intervention while ensuring complete coverage of the relevant examination area.
3Ease of operation
If simple graphics (circle, rectangle) are used to define the examination region, then the setting process is simple, but the examination region cannot correctly match complicated or special shapes
Solution Approach 1:
The patent segments the examination region definition into multiple independent parameters (area ratio, position coordinates, shape constraints) rather than requiring a single geometric shape. This parameter-based segmentation allows the system to flexibly define complex or special shapes by combining multiple constraint parameters, achieving accurate contour matching for complicated objects while maintaining simple automatic setting processes without manual graphic drawing.
Data Source
AI summary
An image examination method includes an acquisition step of acquiring a to-be-examined object image obtained by capturing an image of a to-be-examined object, a setting reading step of reading an examination region definition information from a storage device in which the examination region definition information is previously stored, an examination region extracting step of extracting a portion constituting the examination region as an examination region image from the to-be-examined object image based on the examination region definition information, and an examination process step of examining the to-be-examined object by analyzing the examination region image. The examination region definition information is a parameter used as a constraint condition in an optimum solution search process.