A Comprehensive Augmentation Framework for Anomaly Detection

TL;DR

The paper addresses anomaly detection where real anomalous samples are scarce and class-dependent normality can bias training. It proposes a comprehensive anomaly simulation framework with diverse anomaly types (Transparent and Opaque) and a near-distribution augmentation approach, plus a selective strategy to apply augmentations. It also introduces a split-training protocol for reconstructive models to maintain consistent reconstruction quality and improve generalization. On the MVTec AD benchmark and via a diverse simulated anomaly dataset, the method achieves state-of-the-art performance and shows promise for robust generalization to unseen real-world anomalies.

Abstract

Data augmentation methods are commonly integrated into the training of anomaly detection models. Previous approaches have primarily focused on replicating real-world anomalies or enhancing diversity, without considering that the standard of anomaly varies across different classes, potentially leading to a biased training distribution. This paper analyzes crucial traits of simulated anomalies that contribute to the training of reconstructive networks and condenses them into several methods, thus creating a comprehensive framework by selectively utilizing appropriate combinations. Furthermore, we integrate this framework with a reconstruction-based approach and concurrently propose a split training strategy that alleviates the issue of overfitting while avoiding introducing interference to the reconstruction process. The evaluations conducted on the MVTec anomaly detection dataset demonstrate that our method outperforms the previous state-of-the-art approach, particularly in terms of object classes. To evaluate generalizability, we generate a simulated dataset comprising anomalies with diverse characteristics since the original test samples only include specific types of anomalies and may lead to biased evaluations. Experimental results demonstrate that our approach exhibits promising potential for generalizing effectively to various unforeseen anomalies encountered in real-world scenarios.

Figures (4)

• Figure 1: This figure shows how to select the appropriate augmentations for training.
• Figure 2: This figure illustrates the process of constructing a near-distribution anomaly.
• Figure 3: This figure shows the structure of the reconstructive framework and the concept of the split training strategy. The green samples and the orange samples are from the two different portions of the given training distribution.
• Figure 4: This figure compares the reconstruction quality between DRAEM (second row) and ours (third row).