Problem
Research questions and friction points this paper is trying to address.
Image Analysis
Complex Backgrounds
Anomaly Detection
Ano-SuPs: Multi-size anomaly detection for manufactured products by identifying suspected patches
🤖 AI Summary
Complex backgrounds and multi-scale anomalies in manufacturing images cause modeling bias and anomaly contamination, undermining the reliability of unsupervised defect detection. Method: This paper proposes a two-stage suspicious patch iterative identification framework: first, suspicious patches are filtered and removed via dual-reconstruction residual analysis and statistical consistency verification to construct a clean normal sample set; subsequently, this purified set drives adaptive patch-wise reconstruction and precise anomaly localization. Contribution/Results: The method decouples anomaly interference from model learning, eliminating reliance on prior assumptions about anomaly characteristics—unlike conventional matrix decomposition approaches—and supports detection of arbitrarily sized defects. Evaluated on both synthetic and real-world production-line datasets, it achieves an average 12.6% improvement in F1-score, significantly enhancing detection sensitivity for small and irregular anomalies. Moreover, it demonstrates superior robustness against background clutter and stronger cross-scenario generalization compared to state-of-the-art unsupervised methods.
📝 Abstract
Image-based systems have gained popularity owing to their capacity to provide rich manufacturing status information, low implementation costs and high acquisition rates. However, the complexity of the image background and various anomaly patterns pose new challenges to existing matrix decomposition methods, which are inadequate for modeling requirements. Moreover, the uncertainty of the anomaly can cause anomaly contamination problems, making the designed model and method highly susceptible to external disturbances. To address these challenges, we propose a two-stage strategy anomaly detection method that detects anomalies by identifying suspected patches (Ano-SuPs). Specifically, we propose to detect the patches with anomalies by reconstructing the input image twice: the first step is to obtain a set of normal patches by removing those suspected patches, and the second step is to use those normal patches to refine the identification of the patches with anomalies. To demonstrate its effectiveness, we evaluate the proposed method systematically through simulation experiments and case studies. We further identified the key parameters and designed steps that impact the model's performance and efficiency.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Ano-SuPs
Two-stage Anomaly Detection
Robustness Improvement
🔎 Similar Papers
No similar papers found.
