A Dual-Encoder Framework with Global-to-Local Modulation for Multi-Class Unsupervised Anomaly Detection
Authors:
- Yunxin LiuYLYunxin Liu
- Danqing LiuDLDanqing Liu
- Qinfen WangQWQinfen Wang
- Qian YuQYQian Yu
- Jiaxin LiJLJiaxin Li
- Yanhui GuoYGYanhui Guo
Preprint
- 23 June 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
Reconstruction-based methods leveraging Vision Transformers (ViT) have demonstrated significant potential in unsupervised industrial anomaly detection. However, existing single-stream architectures struggle with an inherent trade-off when processing multi-scale features: while high-resolution features enhance pixel-level localization, they often introduce local noise that degrades image-level detection performance. To resolve this conflict between global semantic stability and local detail sensitivity, we propose a dual-encoder framework named GLM-AD (Global-to-Local Modulation for Anomaly Detection). In which, the global encoder captures stable structural semantics, while the local encoder focuses on fine-grained defect-sensitive patterns. On top of this architecture, we introduce a Global-to-Local Feature Modulation (GLFM) module, which performs channel-wise modulation to implicitly constrain local feature reconstruction using global semantic cues. Furthermore, to cope with the significant morphological heterogeneity of industrial anomalies, we propose a Multi-Scale Ensemble Scoring (MSE-Scoring) strategy. As a model-agnostic and label-free post-processing scheme, MSE-Scoring integrates anomaly responses across multiple observation profiles, effectively alleviating the signal-to-noise imbalance inherent in high-resolution features. Extensive experiments conducted on three public industrial benchmarks(MVTec-AD, VisA, and Real-IAD) demonstrate that GLM-AD consistently achieves state-of-the-art performance in terms of detection robustness and localization accuracy, validating its effectiveness and practical applicability in real-world industrial inspection scenarios.Keywords
This publication has 10 references indexed in Scilit:
- Boosting Fine-Grained Visual Anomaly Detection with Coarse-Knowledge-Aware Adversarial LearningProceedings of the AAAI Conference on Artificial Intelligence, 2025
- RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detection2024
- Deep Industrial Image Anomaly Detection: A SurveyMachine Intelligence Research, 2024
- FastRecon: Few-shot Industrial Anomaly Detection via Fast Feature Reconstruction2023
- Detection of Surface Defects on Railway Tracks Based on Deep LearningIEEE Access, 2022
- A Unified Model for Multi-Class Anomaly Detection2022
- Reconstruction by inpainting for visual anomaly detectionPattern Recognition, 2021
- Segmentation-based deep-learning approach for surface-defect detectionJournal of Intelligent Manufacturing, 2019
- Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders2019
- ImageNet: A large-scale hierarchical image database2009