MAEDAY: MAE for few and zero shot AnomalY-Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Schwartz, Eli, Arbelle, Assaf, Karlinsky, Leonid, Harary, Sivan, Scheidegger, Florian, Doveh, Sivan, Giryes, Raja
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914679312351232
author Schwartz, Eli
Arbelle, Assaf
Karlinsky, Leonid
Harary, Sivan
Scheidegger, Florian
Doveh, Sivan
Giryes, Raja
author_facet Schwartz, Eli
Arbelle, Assaf
Karlinsky, Leonid
Harary, Sivan
Scheidegger, Florian
Doveh, Sivan
Giryes, Raja
contents 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 .
format Preprint
id arxiv_https___arxiv_org_abs_2211_14307
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle MAEDAY: MAE for few and zero shot AnomalY-Detection
Schwartz, Eli
Arbelle, Assaf
Karlinsky, Leonid
Harary, Sivan
Scheidegger, Florian
Doveh, Sivan
Giryes, Raja
Computer Vision and Pattern Recognition
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 .
title MAEDAY: MAE for few and zero shot AnomalY-Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2211.14307