VisA

Visual Anomaly Dataset

Dataset Information
Introduced
2022
License
Unknown
Homepage
Official Website
Contents

Overview

The VisA dataset contains 12 subsets corresponding to 12 different objects as shown in the above figure. There are 10,821 images with 9,621 normal and 1,200 anomalous samples. Four subsets are different types of printed circuit boards (PCB) with relatively complex structures containing transistors, capacitors, chips, etc. For the case of multiple instances in a view, we collect four subsets: Capsules, Candles, Macaroni1 and Macaroni2. Instances in Capsules and Macaroni2 largely differ in locations and poses. Moreover, we collect four subsets including Cashew, Chewing gum, Fryum and Pipe fryum, where objects are roughly aligned. The anomalous images contain various flaws, including surface defects such as scratches, dents, color spots or crack, and structural defects like misplacement or missing parts.

Variants: VisA

Associated Benchmarks

This dataset is used in 2 benchmarks:

Recent Benchmark Submissions

Task Model Paper Date
Anomaly Detection CRAS Center-aware Residual Anomaly Synthesis for … 2025-05-23
Anomaly Detection INP-Former ViT-B (model-unified multi-class) Exploring Intrinsic Normal Prototypes within … 2025-03-04
Anomaly Detection PBAS Progressive Boundary Guided Anomaly Synthesis … 2024-12-23
Anomaly Detection URD Unlocking the Potential of Reverse … 2024-12-10
Anomaly Detection SuperSimpleNet SuperSimpleNet: Unifying Unsupervised and Supervised … 2024-08-06
Anomaly Detection AdaCLIP AdaCLIP: Adapting CLIP with Hybrid … 2024-07-22
Anomaly Detection VCP-CLIP VCP-CLIP: A visual context prompting … 2024-07-17
Anomaly Detection GLASS A Unified Anomaly Synthesis Strategy … 2024-07-12
Anomaly Detection GLAD GLAD: Towards Better Reconstruction with … 2024-06-11
Anomaly Detection AnomalyDINO-S (2-shot) AnomalyDINO: Boosting Patch-based Few-shot Anomaly … 2024-05-23
Anomaly Detection AnomalyDINO-S (full-shot) AnomalyDINO: Boosting Patch-based Few-shot Anomaly … 2024-05-23
Anomaly Detection Dinomaly ViT-L (model-unified multi-class) Dinomaly: The Less Is More … 2024-05-23
Anomaly Detection AnomalyDINO-S (4-shot) AnomalyDINO: Boosting Patch-based Few-shot Anomaly … 2024-05-23
Anomaly Detection AnomalyDINO-S (1-shot) AnomalyDINO: Boosting Patch-based Few-shot Anomaly … 2024-05-23
Anomaly Detection RealNet RealNet: A Feature Selection Network … 2024-03-09
Anomaly Detection MuSc (zero-shot) MuSc: Zero-Shot Industrial Anomaly Classification … 2024-01-30
Anomaly Detection D3AD Dynamic Addition of Noise in … 2024-01-09
Anomaly Detection TransFusion TransFusion -- A Transparency-Based Diffusion … 2023-11-16
Anomaly Detection AnomalyCLIP AnomalyCLIP: Object-agnostic Prompt Learning for … 2023-10-29
Anomaly Detection FAIRnoDTD FAIR: Frequency-aware Image Restoration for … 2023-09-13

Research Papers

Recent papers with results on this dataset:

External Links: