Weighted k-NN Anomaly Detection for Imbalanced Part Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing anomaly detection systems face challenges in accurately distinguishing between good and bad objects due to data imbalance, with existing techniques being inflexible and ineffective in improving detection results.
Innovation Solution
A method involving a feature extractor neural network and a weighted k-nearest neighbors (k-NN) model with gradient ascent to iteratively update filter weights, enhancing separation between good and bad scores through threshold and model adaptation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional k-NN anomaly detection is used with equal filter weights, then the system is simple to implement, but it cannot effectively distinguish between good and bad objects due to data imbalance
Solution Approach 1:
The patent applies parameter changes by introducing filter weights as adjustable parameters in the k-NN anomaly detection system. Instead of using equal weights for all filters, the system optimizes weight values to enhance the separation between good and bad object scores, directly addressing the data imbalance problem while maintaining the fundamental k-NN approach
Solution Approach 2:
The patent implements dynamics by making the filter weights adaptive rather than static. The weights are optimized through iterative processes that adjust them based on the data characteristics, allowing the system to dynamically adapt to the specific anomaly detection task and improve discrimination between normal and anomalous objects
2Reliability
If filter weights are optimized to increase separation between good and bad scores, then recall and precision improve, but the computational complexity increases
Solution Approach 1:
The patent employs feedback mechanisms where the anomaly detection system uses the scores from good and bad objects to guide the optimization of filter weights. The weight optimization process receives feedback from the detection performance and adjusts weights iteratively to maximize the separation between classes, thereby improving reliability through a closed-loop optimization process
Solution Approach 2:
The patent applies preliminary action by pre-optimizing the filter weights before actual anomaly detection is performed. The weight optimization is conducted in advance using training data, creating a prepared model that can then be deployed for efficient real-time detection without requiring complex computations during the actual detection phase
Data Source
AI summary
A method and system for anomaly detection from object images. A feature extractor provides feature data characterizing part images, including good parts and bad parts. Training data for good parts is used to create a k-nearest neighbors (k-NN) model core set. The feature data includes hundreds of feature vectors, each having hundreds of filter dimensions. A weight value of one is initially assigned to each filter, and test data comprising some good and some bad parts is evaluated by a weighted k-NN module to determine an anomaly score from the weighted feature data. After all test images are evaluated, good and bad data points nearest a threshold are selected and a gradient ascent computation is performed to update the filter weights. Anomaly scoring and gradient ascent are performed iteratively until filter weights are identified which maximize the separation between good and bad scores, thereby eliminating missed detections and false anomalies.