Home Console Docs News Papers Tutorials Datasets Wiki SOTA LLM Models GPU Leaderboard Events

English

Anomaly Detection

Anomaly detection is a binary classification task aimed at identifying unusual or unexpected patterns that significantly deviate from the majority of data in a dataset. The goal of this task is to discover these outliers, which may represent errors, fraud, or other types of abnormal events, and mark them for further investigation. Anomaly detection has significant application value in areas such as financial risk control, cybersecurity, and medical diagnosis.

One-class CIFAR-10

UCR Anomaly Archive

Auto-Encoder with Regression (AER)

Fishyscapes L&F

One-class CIFAR-100

MuSc (zero-shot)

Background-Agnostic

Unlabeled CIFAR-10 vs CIFAR-100

One-class ImageNet-30

Numenta Anomaly Benchmark

Anomaly Detection on Unlabeled CIFAR-10 vs LSUN (Fix)

Leave-One-Class-Out ImageNet-30

Leave-One-Class-Out CIFAR-10

Hyper-Kvasir Dataset

AttentDifferNet (SENet-AlexNet)

Anomaly Detection on Anomaly Detection on Unlabeled ImageNet-30 vs Flowers-102

Anomaly Detection on Unlabeled ImageNet-30 vs CUB-200

MVTEC AD textures

Surface Defect Saliency of Magnetic Tile

Self-Supervised One-class SVM, RBF kernel

Self-Supervised One-class SVM, RBF kernel

UEA time-series datasets

MVTec AD Textures Domain Generalization

ASSIRA Cat Vs Dog

Shell-based Anomaly (supervisered)

MVTec 3D-AD (RGB)

PCA via oversampling

Forest CoverType

ShanghaiTech Campus

MIT-BIH Arrhythmia Database

Kaggle-Credit Card Fraud Dataset

Shell-based Anomaly (supervised)

Anomaly Detection | SOTA | HyperAI