Recognition Image Parameter Determination for Accuracy and Viewability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Images captured by imaging devices are often optimized for human visibility but may not be optimal for recognition by recognition apparatuses, leading to suboptimal recognition performance.
Innovation Solution
A parameter determination apparatus and method that calculates and determines image generation parameters based on recognition results to optimize images for better recognition by recognition apparatuses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If images are optimized for human visibility by adjusting imaging parameters, then the images are easy to be viewed by persons, but the recognition apparatus cannot properly recognize the image
Solution Approach 1:
The patent changes the imaging parameters (optical characteristics and image processing parameters) from those optimized for human viewing to those optimized for recognition apparatus processing. This involves adjusting parameters such as color space, brightness, contrast, and other optical characteristics to create images that facilitate accurate recognition while maintaining reasonable viewability.
Solution Approach 2:
The patent dynamically adjusts imaging parameters based on the specific recognition task and conditions. Rather than using fixed parameters, the system adapts the optical characteristics and processing parameters according to the recognition requirements, enabling optimal balance between viewability and recognition accuracy for different scenarios.
2Measurement precision
If imaging parameters are adjusted to improve recognition performance, then the recognition apparatus can properly recognize the image, but the image is no longer easy to be viewable by a person
Solution Approach 1:
The patent applies parameter changes to find an optimal balance point where recognition accuracy is significantly improved while viewability degradation is minimized. This involves carefully adjusting optical characteristics and image processing parameters to achieve the necessary recognition performance without making the image completely unsuitable for human viewing.
Solution Approach 2:
The patent applies partial optimization to specific image regions or parameters that are most critical for recognition, rather than uniformly optimizing all parameters. This allows selective adjustment of certain optical characteristics or processing parameters that provide the most benefit for recognition while maintaining better overall viewability.
3Measurement precision
If multiple image processing parameters are adjusted to optimize recognition, then the recognition accuracy improves, but the device complexity increases
Solution Approach 1:
The patent segments the image processing into distinct stages with specific parameters optimized for each stage. Rather than adjusting all parameters simultaneously, the system divides the processing into separate steps (e.g., optical characteristic adjustment, image processing parameter adjustment, recognition-specific optimization), making the complex parameter adjustment more manageable and systematic.
Solution Approach 2:
The patent performs preliminary determination of optimal imaging parameters based on the recognition task requirements before actual image capture or processing. This advance preparation of parameter settings reduces the complexity of real-time adjustments and enables more systematic optimization of recognition accuracy without excessive device complexity.
Data Source
AI summary
A parameter determination apparatus (3) includes: a calculation unit (313) that is configured to calculate, based on a recognized result of a plurality of recognition target images by a recognition apparatus (2) that performs a recognition operation on the recognition target image (100, 200), an evaluation value for evaluating the recognized result; and a determination unit (314) that is configured to determine, based on the evaluation value, an image generation parameter (300, 301, 302 303b) that is used to generate the recognition target image.