ZSG-IAD: A Multimodal Framework for Zero-Shot Grounded Industrial Anomaly Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Qiuhui, Song, Jiaxiang, Tan, Shuai, Zhong, Weimin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915945475211264
author Chen, Qiuhui
Song, Jiaxiang
Tan, Shuai
Zhong, Weimin
author_facet Chen, Qiuhui
Song, Jiaxiang
Tan, Shuai
Zhong, Weimin
contents Deep learning-based industrial anomaly detectors often behave as black boxes, making it hard to justify decisions with physically meaningful defect evidence. We propose ZSG-IAD, a multimodal vision-language framework for zero-shot grounded industrial anomaly detection. Given RGB images, sensor images, and 3D point clouds, ZSG-IAD generates structured anomaly reports and pixel-level anomaly masks. ZSG-IAD introduces a language-guided two-hop grounding module: (1) anomaly-related sentences select evidence-like latent slots distilled from multimodal features, yielding coarse spatial support; (2) selected slots modulate feature maps via channel-spatial gating and a lightweight decoder to produce fine-grained masks. To improve reliability, we further apply Executable-Rule GRPO with verifiable rewards to promote structured outputs, anomaly-region consistency, and reasoning-conclusion coherence. Experiments across multiple industrial anomaly benchmarks show strong zero-shot performance and more transparent, physically grounded explanations than prior methods. We will release code and annotations to support future research on trustworthy industrial anomaly detection systems.
format Preprint
id arxiv_https___arxiv_org_abs_2604_17949
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ZSG-IAD: A Multimodal Framework for Zero-Shot Grounded Industrial Anomaly Detection
Chen, Qiuhui
Song, Jiaxiang
Tan, Shuai
Zhong, Weimin
Computer Vision and Pattern Recognition
Deep learning-based industrial anomaly detectors often behave as black boxes, making it hard to justify decisions with physically meaningful defect evidence. We propose ZSG-IAD, a multimodal vision-language framework for zero-shot grounded industrial anomaly detection. Given RGB images, sensor images, and 3D point clouds, ZSG-IAD generates structured anomaly reports and pixel-level anomaly masks. ZSG-IAD introduces a language-guided two-hop grounding module: (1) anomaly-related sentences select evidence-like latent slots distilled from multimodal features, yielding coarse spatial support; (2) selected slots modulate feature maps via channel-spatial gating and a lightweight decoder to produce fine-grained masks. To improve reliability, we further apply Executable-Rule GRPO with verifiable rewards to promote structured outputs, anomaly-region consistency, and reasoning-conclusion coherence. Experiments across multiple industrial anomaly benchmarks show strong zero-shot performance and more transparent, physically grounded explanations than prior methods. We will release code and annotations to support future research on trustworthy industrial anomaly detection systems.
title ZSG-IAD: A Multimodal Framework for Zero-Shot Grounded Industrial Anomaly Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.17949