paper-with-me

Papers

Statistical Non-linear Reconstruction Loss for Image Anomaly Detection

2026-07-14 · Nguyen Minh Tri, Hoang Khuong Duy, Huynh Cong Viet Ngu arxiv

Reconstruction-based methods are a cornerstone of unsupervised image anomaly detection, but they remain vulnerable to \emph{outlier leakage}, where standard mean squared error (MSE) loss drives the model to faithfully reconstruct anomalous patterns. We propose a Non-linear Reconstruction Loss that applies a sigmoid-based squashing function to suppress high-magnitude features, preventing outliers from dominating optimization while preserving sensitivity to normal patterns. In addition, we introduce a statistical calibration scheme that selects the scaling factor $k$ from the confidence interval (CI) of the normal feature distribution, enabling data-driven control of the suppression strength. Our approach achieves competitive or superior anomaly detection performance compared to state-of-the-art methods, reaching 99.0\% Image-AUROC and 97.3\% Pixel-AUROC on MVTec-AD, and 95.3\% Image-AUROC and 99.0\% Pixel-AUROC on VisA. These results indicate that non-linear gradient suppression is an effective mechanism for mitigating outlier leakage and improving anomaly localization in unified industrial inspection settings. The implementation is available at https://github.com/mintii13/Statistical-Non-linear-Reconstruction-Loss.git.

📄 PDF Abstract BibTeX arXiv:2607.12866

Code (0)

등록된 구현이 없습니다.

Tasks

Anomaly Detection

Results from the Paper

RankTaskDatasetModelMetrics
#4 Anomaly Detection VisA Non-linear Detection AUROC: 99.0

Similar Papers 제목 키워드 기반

DDR-ID: Dual Deep Reconstruction Networks Based Image Decomposition for Anomaly Detection

2020-07-18 · Dongyun Lin, Yiqun Li, Shudong Xie, Tin Lay Nwe 외

One pivot challenge for image anomaly (AD) detection is to learn discriminative information only from normal class training images. Most image reconstruction based AD methods rely on the discriminative capability of reco…

Adversarial AttackAdversarial Attack DetectionAnomaly DetectionBenchmarking+1

Anomaly Detection By Autoencoder Based On Weighted Frequency Domain Loss

2021-05-21 · Masaki Nakanishi, Kazuki Sato, Hideo Terada

In image anomaly detection, Autoencoders are the popular methods that reconstruct the input image that might contain anomalies and output a clean image with no abnormalities. These Autoencoder-based methods usually calcu…

Anomaly Detection

COGNOS: Universal Enhancement for Time Series Anomaly Detection via Constrained Gaussian-Noise Optimization and Smoothing

2025-11-10 · Wenlong Shang, Shihao Tian, Xutong Wan, Peng Chang arxiv

Reconstruction-based methods are a dominant paradigm in time series anomaly detection (TSAD), however, their near-universal reliance on Mean Squared Error (MSE) loss results in statistically flawed reconstruction residua…

Time Series Anomaly Detection

FRE: A Fast Method For Anomaly Detection And Segmentation

2022-11-23 · Ibrahima Ndiour, Nilesh Ahuja, Utku Genc, Omesh Tickoo

This paper presents a fast and principled approach for solving the visual anomaly detection and segmentation problem. In this setup, we have access to only anomaly-free training data and want to detect and identify anoma…

Anomaly DetectionCPUDimensionality Reduction

Autoencoders for Anomaly Detection are Unreliable

2025-01-23 · Roel Bouman, Tom Heskes

Autoencoders are frequently used for anomaly detection, both in the unsupervised and semi-supervised settings. They rely on the assumption that when trained using the reconstruction loss, they will be able to reconstruct…

Anomaly Detection