MAEDAY: MAE for few and zero shot AnomalY-Detection
This enables anomaly detection with minimal or no normal samples, addressing a practical challenge in industrial inspection and medical imaging.
The paper tackled anomaly detection in images by using a Masked Auto-Encoder (MAE) pre-trained on image inpainting, achieving strong performance for few-shot and zero-shot tasks, including zero-shot foreign object detection.
We propose using Masked Auto-Encoder (MAE), a transformer model self-supervisedly trained on image inpainting, for anomaly detection (AD). Assuming anomalous regions are harder to reconstruct compared with normal regions. MAEDAY is the first image-reconstruction-based anomaly detection method that utilizes a pre-trained model, enabling its use for Few-Shot Anomaly Detection (FSAD). We also show the same method works surprisingly well for the novel tasks of Zero-Shot AD (ZSAD) and Zero-Shot Foreign Object Detection (ZSFOD), where no normal samples are available. Code is available at https://github.com/EliSchwartz/MAEDAY .