Spot The Odd One Out: Regularized Complete Cycle Consistent Anomaly Detector GAN

Zahra Dehghanian, Saeed Saravani, Maryam Amirmazlaghani, Mohammad Rahmati
4/16/2023
cs.LG

Official code: ZahraDehghanian97/RCALAD (14 stars, MIT, last commit 2024-12-31).

Abstract

This study presents an adversarial method for anomaly detection in real-world applications, leveraging the power of generative adversarial neural networks (GANs) through cycle consistency in reconstruction error. Previous methods suffer from the high variance between class-wise accuracy which leads to not being applicable for all types of anomalies. The proposed method named RCALAD tries to solve this problem by introducing a novel discriminator to the structure, which results in a more efficient training process. Additionally, RCALAD employs a supplementary distribution in the input space to steer reconstructions toward the normal data distribution, effectively separating anomalous samples from their reconstructions and facilitating more accurate anomaly detection. To further enhance the performance of the model, two novel anomaly scores are introduced. The proposed model has been thoroughly evaluated through extensive experiments on six various datasets, yielding results that demonstrate its superiority over existing state-of-the-art models. The code is readily available to the research community at https://github.com/zahraDehghanian97/RCALAD.

View on arXivView PDF

Code Implementations(1)

ZahraDehghanian97/RCALADOfficial100%
MIT

Implementation of RCALAD Model based on "Anomaly Detection Using Complete Cycle Consistent Generative Adversarial Network" paper *Official Code

140PythonApr 15, 20201 years agoMIT
adversarial-learninganomaly-detectionganneural-networkrcalad

Similar Papers

TSI-GAN: Unsupervised Time Series Anomaly Detection using Convolutional Cycle-Consistent Generative Adversarial Networks

84% match1 repo122023

Adversarial Anomaly Detection using Gaussian Priors and Nonlinear Anomaly Scores

83% match2023

Collaborative Reconstruction and Repair for Multi-class Industrial Anomaly Detection

82% match2025

AGAD: Adversarial Generative Anomaly Detection

82% match2023

Anomaly Detection in Power Generation Plants with Generative Adversarial Networks

82% match2023

Unsupervised Anomaly Detection and Localization of Machine Audio: A GAN-based Approach

81% match1 repo522023

Cite this paper

@article{dehghanian2023spot,
  title  = {Spot The Odd One Out: Regularized Complete Cycle Consistent Anomaly Detector GAN},
  author = {Zahra Dehghanian and Saeed Saravani and Maryam Amirmazlaghani and Mohammad Rahmati},
  year   = {2023},
  eprint = {2304.07769},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2304.07769v3}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2304.07769.svg)](https://codeofpaper.com/paper/2304.07769)

Discussion