Visual Prompting Meets Feature Reconstruction-Based Anomaly Detection with Dual-Teacher Supervision

2026-06-08

AI Summary

This work addresses the significant performance degradation of existing anomaly detection methods under variations in object scale, viewpoint, background, and illumination, which hinders their applicability in real-world complex scenarios. To overcome this limitation, the authors propose a novel framework that integrates visual prompting with feature reconstruction. The approach isolates the target region using a foreground-background mask, employs a tunable dual-teacher supervision mechanism to enhance domain adaptability, and leverages a diffusion model to generate synthetic data for effective augmentation. Evaluated on the AeBAD dataset, the method achieves state-of-the-art performance, surpassing prior approaches by 3.5 percentage points, and demonstrates superior robustness and generalization capability in both anomaly detection and segmentation tasks.

Abstract

Recent Anomaly Detection methods achieve perfect detection and segmentation scores on well-established datasets, such as MVTec. However, many of these methods face challenges when foundational assumptions - such as consistent object scale, viewpoint, background, illumination, and centered placement - are violated. Those variations that occur render anomaly detection methods unusable in many real-world scenarios. To address these limitations, we introduce three key contributions: (1) a visual prompting pipeline that isolates objects using foreground-background masking; (2) a mechanism for unfreezing the teacher in student-teacher models to improve domain adaptability; and (3) a data augmentation strategy leveraging diffusion-generated synthetic images to enhance anomaly detection performance. We achieve a 3.5 percentage point improvement over the previous state-of-the-art on the challenging AeBAD dataset by using the Masked Multiscale Reconstruction (MMR) model as our backbone.

Problem

• Anomaly Detection
• Domain Adaptation
• Real-world Scenarios
• Object Variability
• Visual Prompting

Innovation

• visual prompting
• dual-teacher supervision
• feature reconstruction
• diffusion-based augmentation
• anomaly detection