High-Fidelity Zero-Shot Texture Anomaly Localization Using Feature Correspondence Analysis

04/13/2023

∙

We propose a novel method for Zero-Shot Anomaly Localization that leverages a bidirectional mapping derived from the 1-dimensional Wasserstein Distance. The proposed approach allows pinpointing the anomalous regions in a texture with increased precision by aggregating the contribution of a pixel to the errors of all nearby patches. We validate our solution on several datasets and obtain more than a 40 MVTec AD dataset in a zero-shot setting.

READ FULL TEXT

High-Fidelity Zero-Shot Texture Anomaly Localization Using Feature Correspondence Analysis

Zero-Shot Anomaly Detection without Foundation Models

AnoVL: Adapting Vision-Language Models for Unified Zero-shot Anomaly Localization

Segment Any Anomaly without Training via Hybrid Prompt Regularization

Winning Solution for the CVPR2023 Visual Anomaly and Novelty Detection Challenge: Multimodal Prompting for Data-centric Anomaly Detection

A Zero-/Few-Shot Anomaly Classification and Segmentation Method for CVPR 2023 VAND Workshop Challenge Tracks 1 2: 1st Place on Zero-shot AD and 4th Place on Few-shot AD

WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation

Edge Storage Management Recipe with Zero-Shot Data Compression for Road Anomaly Detection

High-Fidelity Zero-Shot Texture Anomaly Localization Using Feature Correspondence Analysis

Related Research

Zero-Shot Anomaly Detection without Foundation Models

AnoVL: Adapting Vision-Language Models for Unified Zero-shot Anomaly Localization

Segment Any Anomaly without Training via Hybrid Prompt Regularization

Winning Solution for the CVPR2023 Visual Anomaly and Novelty Detection Challenge: Multimodal Prompting for Data-centric Anomaly Detection

A Zero-/Few-Shot Anomaly Classification and Segmentation Method for CVPR 2023 VAND Workshop Challenge Tracks 1 2: 1st Place on Zero-shot AD and 4th Place on Few-shot AD

WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation

Edge Storage Management Recipe with Zero-Shot Data Compression for Road Anomaly Detection