Process Cycle Image Classification for Genotyping Root Cause Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing genotyping processes are vulnerable to mechanical and chemical errors, leading to low-quality images that do not provide insight into the root cause of failure, and current quality scoring methods fail to distinguish between successful and unsuccessful production runs.
Innovation Solution
A machine learning-based approach using Principal Component Analysis (PCA) and image thresholding techniques to classify images and determine root causes of failure in genotyping processes, employing classifiers such as random forest to analyze images from genotyping instruments, reducing image resolution for computational efficiency, and extracting features from active areas to identify failure categories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional quality scoring methods are used to evaluate genotyping process images, then the evaluation process is simple and fast, but the method cannot provide insight into the root cause of failure and cannot distinguish between successful and unsuccessful production runs
Solution Approach 1:
The patent segments the image analysis into multiple specialized components: a binary classifier that separates successful from unsuccessful runs, and a root cause classifier that identifies specific failure types. This segmentation allows the system to provide detailed root cause information without requiring a single complex analysis system, thereby resolving the contradiction between information completeness and system complexity.
Solution Approach 2:
The patent introduces image feature extraction as an intermediary step between raw image input and classification. By extracting relevant features (such as signal intensity, spatial distribution, and morphological characteristics) before classification, the system can provide comprehensive root cause analysis while managing complexity through feature selection and dimensionality reduction techniques.
2Measurement precision
If full-resolution images are analyzed to ensure complete information is available, then measurement precision is improved, but computational resources and processing time increase significantly
Solution Approach 1:
The patent extracts and analyzes only the most relevant image features rather than processing complete high-resolution images. By taking out and focusing on discriminative features (such as key signal patterns and critical region characteristics), the system maintains measurement precision while significantly reducing computational resources and processing time.
Solution Approach 2:
The patent applies partial action by using reduced-resolution images or image patches for initial classification and root cause identification. This approach processes only the necessary portion of the image data required to achieve accurate classification, avoiding the excessive computational burden of analyzing entire high-resolution images while maintaining sufficient precision for the task.
3Loss of information
If detailed root cause analysis is performed on all images, then information completeness is improved, but processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary classification to identify unsuccessful production runs before conducting detailed root cause analysis. This two-stage approach ensures that comprehensive failure analysis information is generated only when necessary (for failed runs), while successful runs are quickly identified and processed with minimal computational overhead, thereby maintaining both information completeness and processing throughput.
Solution Approach 2:
The patent applies partial action by performing detailed root cause analysis only on images that are classified as unsuccessful, rather than analyzing all images in detail. This selective approach provides complete failure analysis information where needed while avoiding unnecessary processing of successful runs, thus resolving the contradiction between information completeness and processing productivity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The technology disclosed relates to classification of process cycle images to predict success or failure of process cycles. The technology disclosed includes capturing and processing images of sections arranged on an image generating chip in genotyping process. Image description features of production cycle images are created and given as input to classifiers. A trained classifier separates successful production images from unsuccessful or failed production images. The failed production images are further classified by a trained root cause classifier into various categories of failure.