Dual-path Frequency Discriminators for Few-shot Anomaly Detection
This addresses the problem of detecting inconspicuous anomalies with limited normal samples in industrial settings, representing an incremental improvement over existing methods.
The paper tackles few-shot anomaly detection in industrial manufacturing by proposing a dual-path frequency discriminators network that transforms images to the frequency domain to make subtle anomalies more noticeable, achieving state-of-the-art results on MVTec AD and VisA benchmarks.
Few-shot anomaly detection (FSAD) plays a crucial role in industrial manufacturing. However, existing FSAD methods encounter difficulties leveraging a limited number of normal samples, frequently failing to detect and locate inconspicuous anomalies in the spatial domain. We have further discovered that these subtle anomalies would be more noticeable in the frequency domain. In this paper, we propose a Dual-Path Frequency Discriminators (DFD) network from a frequency perspective to tackle these issues. The original spatial images are transformed into multi-frequency images, making them more conducive to the tailored discriminators in detecting anomalies. Additionally, the discriminators learn a joint representation with forms of pseudo-anomalies. Extensive experiments conducted on MVTec AD and VisA benchmarks demonstrate that our DFD surpasses current state-of-the-art methods. The code is available at \url{https://github.com/yuhbai/DFD}.