Do LLMs Understand Visual Anomalies? Uncovering LLM's Capabilities in Zero-shot Anomaly Detection

Jiaqi Zhu, Shaofeng Cai*, Fang Deng*, Beng Chin Ooi, Junran Wu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Citations (Scopus)

Abstract

Large vision-language models (LVLMs) are markedly proficient in deriving visual representations guided by natural language. Recent explorations have utilized LVLMs to tackle zero-shot visual anomaly detection (VAD) challenges by pairing images with textual descriptions indicative of normal and abnormal conditions, referred to as anomaly prompts. However, existing approaches depend on static anomaly prompts that are prone to cross-semantic ambiguity, and prioritize global image-level representations over crucial local pixel-level image-to-text alignment that is necessary for accurate anomaly localization. In this paper, we present ALFA, a training-free approach designed to address these challenges via a unified model. We propose a run-time prompt adaptation strategy, which first generates informative anomaly prompts to leverage the capabilities of a large language model (LLM). This strategy is enhanced by a contextual scoring mechanism for per-image anomaly prompt adaptation and cross-semantic ambiguity mitigation. We further introduce a novel fine-grained aligner to fuse local pixel-level semantics for precise anomaly localization, by projecting the image-text alignment from global to local semantic spaces. Extensive evaluations on the challenging MVTec and VisA datasets confirm ALFA's effectiveness in harnessing the language potential for zero-shot VAD, achieving significant PRO improvements of 12.1% on MVTec AD and 8.9% on VisA compared to state-of-the-art zero-shot VAD approaches.

Original languageEnglish
Title of host publicationMM 2024 - Proceedings of the 32nd ACM International Conference on Multimedia
PublisherAssociation for Computing Machinery, Inc
Pages48-57
Number of pages10
ISBN (Electronic)9798400706868
DOIs
Publication statusPublished - 28 Oct 2024
Event32nd ACM International Conference on Multimedia, MM 2024 - Melbourne, Australia
Duration: 28 Oct 20241 Nov 2024

Publication series

NameMM 2024 - Proceedings of the 32nd ACM International Conference on Multimedia

Conference

Conference32nd ACM International Conference on Multimedia, MM 2024
Country/TerritoryAustralia
CityMelbourne
Period28/10/241/11/24

Keywords

  • industrial manufacturing
  • large vision-language model
  • visual anomaly detection
  • zero-shot detection

Fingerprint

Dive into the research topics of 'Do LLMs Understand Visual Anomalies? Uncovering LLM's Capabilities in Zero-shot Anomaly Detection'. Together they form a unique fingerprint.

Cite this