Inspection Model Feedback Loop for Defect Classification Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current inspection systems using machine learning models for determining product quality often face challenges in accurately distinguishing between defective and non-defective products, leading to potential errors in classification, which can result in defective products being mistakenly identified as non-defective and shipped.
Innovation Solution
The system incorporates machine learning circuitry for initial classification, followed by a secondary evaluation using feature data acquisition and parameter update mechanisms to refine the learning parameters based on reevaluated data, ensuring accurate determination and reducing erroneous classifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If machine learning circuitry is used to determine whether objects belong to a predetermined attribute, then productivity is improved through automated inspection, but measurement precision deteriorates due to potential classification errors
Solution Approach 1:
The inspection process is segmented into multiple stages: initial machine learning-based classification followed by selective reevaluation of excluded objects. This segmentation allows the system to maintain high productivity for clearly classified objects while applying additional verification only where needed, thus improving measurement precision without sacrificing overall inspection efficiency.
Solution Approach 2:
The system implements feedback mechanisms where determination results from the machine learning circuitry are used to identify excluded objects that require reevaluation. The feature data acquisition circuitry then collects additional data on these excluded objects, and the parameter update circuitry uses this feedback to refine the machine learning model, creating a continuous improvement loop that enhances classification accuracy over time.
2Measurement precision
If detailed inspection is performed on excluded objects to improve quality determination accuracy, then measurement precision is improved, but loss of time increases due to additional inspection steps
Solution Approach 1:
Instead of performing detailed inspection on all objects, the system applies partial action by conducting reevaluation only on excluded objects that require further verification. This selective approach ensures measurement precision is improved for critical cases while minimizing the overall time loss by avoiding redundant inspection of objects that were clearly classified by the machine learning model.
3Measurement precision
If learning parameters are updated based on reevaluated data, then measurement precision is improved through iterative learning, but device complexity increases due to additional circuitry
Solution Approach 1:
The parameter update circuitry is designed to serve multiple functions: it not only updates learning parameters based on reevaluated data but also integrates with the existing machine learning circuitry and feature data acquisition systems. This multi-functionality allows the system to improve measurement precision without requiring entirely separate verification systems, thereby limiting the increase in device complexity.
Data Source
AI summary
An inspection system includes machine learning circuitry configured to determine whether each of objects belongs to a predetermined attribute based on feature data of each of the objects, feature data acquisition circuitry configured to acquire feature data of reevaluated objects which are determined to belong to the predetermined attribute without using the machine learning circuitry among excluded objects which are determined not to belong to the predetermined attribute by the machine learning circuitry, and parameter update circuitry configured to update a learning parameter of the machine learning circuitry based on teaching data including the acquired feature data acquired by the feature data acquisition circuitry.