Inspection Image Classification with Dynamic Reliability Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The accuracy of classification or determination results can vary when using multiple classifiers or determining sections due to differences in images, leading to suboptimal weight settings that affect the final determination outcome.
Innovation Solution
An information processing device that determines reliability for each determining section's output and uses these reliabilities along with the determination results to derive a comprehensive final determination, incorporating a reliability determining section and a comprehensive determination section.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If weights are determined in advance for each classifier, then the classification process is simplified and can be executed efficiently, but the classification accuracy deteriorates because the fixed weights cannot adapt to variations in different images
Solution Approach 1:
The patent applies the dynamics principle by transitioning from static, pre-determined weights to dynamic, image-specific weights. The reliability determining section calculates weights based on the actual characteristics of each input image, allowing the system to adapt to variations in image quality, content, and conditions. This dynamic weight adjustment resolves the contradiction by maintaining ease of implementation through automated calculation while improving classification accuracy through adaptive weighting.
Solution Approach 2:
The patent implements parameter changes by modifying the weight parameters based on image-specific characteristics. Instead of using fixed weight values, the system changes the weight parameters dynamically according to the reliability assessment of each image. This allows the same classifier framework to be applied across different images with optimized weights, resolving the contradiction between implementation simplicity and classification accuracy.
2Reliability
If multiple classifiers are used to improve classification accuracy, then the determination reliability improves, but the system complexity increases due to the need to integrate and weight multiple classifier outputs
Solution Approach 1:
The patent applies the feedback principle by introducing a reliability determining section that assesses the quality and characteristics of each input image. This feedback mechanism provides information about image reliability that is used to adjust the weights of classifier outputs. The feedback loop allows the system to maintain high determination reliability through multiple classifiers while managing complexity by using the reliability information to optimally combine classifier results.
Solution Approach 2:
The reliability determining section acts as an intermediary between the multiple classifiers and the final determination. It processes the raw classifier outputs and image characteristics, transforming them into weighted contributions that can be efficiently combined. This intermediary component manages the complexity of integrating multiple classifiers while preserving their individual strengths, resolving the contradiction between reliability improvement and system complexity.
3Speed
If fixed weights are applied to classifiers, then the processing speed is maintained, but the determination accuracy deteriorates when image characteristics vary
Solution Approach 1:
The patent applies preliminary action by pre-establishing the framework and algorithms for reliability assessment and dynamic weight calculation. Although the actual weight values are computed for each image, the computational methodology and assessment criteria are predetermined. This allows the system to maintain processing speed through efficient, pre-planned computational paths while achieving accurate, image-specific weight adjustment that improves determination accuracy.
Data Source
AI summary
A final determination result is derived in accordance with target data, in consideration of determination results given by determining sections. An information processing device includes: a reliability determining section that determines, in accordance with an inspection image, reliabilities of determination results given by determining sections each configured to determine a given determination matter in accordance with the inspection image; and a comprehensive determination section configured to determine the given determination matter with use of the determination results and the reliabilities.