P-WAE: Generalized Patch-Wasserstein Autoencoder for Anomaly Screening

TL;DR

This paper introduces P-WAE, a novel autoencoder architecture that enhances anomaly detection by increasing latent space expressiveness and balancing reconstruction quality using sliced-Wasserstein distance, validated on industrial datasets.

Contribution

The paper proposes a patch-wise variational inference model with jigsaw puzzle solving and employs sliced-Wasserstein distance to improve anomaly detection performance.

Findings

Outperforms existing methods on MVTec AD dataset

Enhances latent space expressiveness for high-dimensional data

Balances reconstruction fidelity and representation generalization

Abstract

Anomaly detection plays a pivotal role in numerous real-world scenarios, such as industrial automation and manufacturing intelligence. Recently, variational inference-based anomaly analysis has attracted researchers' and developers' attention. It aims to model the defect-free distribution so that anomalies can be classified as out-of-distribution samples. Nevertheless, there are two disturbing factors that need us to prioritize: (i) the simplistic prior latent distribution inducing limited expressive capability; (ii) the strong probability distance notion results in collapsed features. In this paper, we propose a novel Patch-wise Wasserstein AutoEncoder (P-WAE) architecture to alleviate those challenges. In particular, a patch-wise variational inference model coupled with solving the jigsaw puzzle is designed, which is a simple yet effective way to increase the expressiveness of the latent manifold. This makes using the model on high-dimensional practical data possible. In addition, we leverage a weaker measure, sliced-Wasserstein distance, to achieve the equilibrium between the reconstruction fidelity and generalized representations. Comprehensive experiments, conducted on the MVTec AD dataset, demonstrate the superior performance of our proposed method.