UniSLAD: A Unified Framework for Structural and Logical Industrial Visual Anomaly Detection

2026-06-18

AI Summary

This work addresses the challenge in industrial visual anomaly detection where structural defects and logical anomalies often coexist, yet existing methods struggle to detect both simultaneously. The authors propose UniSLAD, a unified framework capable of jointly identifying these two anomaly types without requiring additional training, making it well-suited for dynamic industrial environments. UniSLAD innovatively integrates CNN and Transformer backbones and introduces a dual-granularity feature representation module to collaboratively model local textures and global context. It further incorporates dual feature extractors, a Mahalanobis distance-enhanced memory bank, and novel pooling mechanisms—Lower-Upper Mean (LUM) and Power Mean Pooling (PMP)—to construct multi-level anomaly representations. Evaluated on two industrial benchmarks, UniSLAD achieves state-of-the-art detection accuracies of 99.4% and 93.1%, significantly outperforming current approaches.

Abstract

Visual anomaly detection is a fundamental task in industrial automation. While existing approaches have achieved notable progress in identifying structural defects, the detection of logical anomalies remains relatively underexplored. In practice, structural and logical anomalies frequently co-occur in industrial workflows. Therefore, a solution capable of detecting both structural and logical anomalies is crucial for advancing comprehensive anomaly detection research. To address this limitation, we propose a unified framework, termed UniSLAD, which jointly addresses logical and structural anomalies without additional training, enabling a practical solution for dynamic industrial environments. First, we introduce a dual-feature extractor that synergistically integrates a Convolutional Neural Network (CNN) backbone for local texture perception with a Transformer backbone for global contextual reasoning, yielding richer and more comprehensive representations. Building on this foundation, we design dual-granularity feature representation modules. At the patch level, memory banks enhanced by the Mahalanobis Transform (MT) preserve representative features and support more discriminative anomaly scoring. At the image level, distribution maps are aggregated using Lower-Upper Mean (LUM) and Power Mean Pooling (PMP), yielding a more robust global representation than conventional average pooling. Extensive experiments on the two industrial benchmarks demonstrate that UniSLAD achieves competitive performance in comprehensive anomaly detection, achieving 99.4% and 93.1%, respectively. Furthermore, ablation studies verify the individual contributions and effectiveness of each proposed component.

Problem

• visual anomaly detection
• structural anomalies
• logical anomalies
• industrial automation
• comprehensive anomaly detection

Innovation

• Unified Anomaly Detection
• Dual-Feature Extractor
• Mahalanobis Transform
• Power Mean Pooling
• Structural and Logical Anomalies