LogicQA: Logical Anomaly Detection with Vision Language Model Generated Questions

2025-03-26

AI Summary

This paper addresses the challenging problem of detecting “visually normal but logically anomalous” instances in industrial process control—such as missing, misaligned, or incorrectly counted objects—without requiring training data or pixel-/image-level annotations. Methodologically, it introduces a novel, interpretable, zero-shot logical anomaly detection framework that leverages vision-language models (VLMs) to automatically generate logic-constraint-oriented questions and convert them into verifiable checklists; these are then evaluated via few-shot reasoning to enable end-to-end logical consistency verification. Its core contribution is the first VLM-driven zero-shot logical constraint modeling paradigm, uniquely balancing high interpretability with practical deployability. Evaluated on the MVTec LOCO AD benchmark, the method achieves state-of-the-art performance (AUROC: 87.6%, F1-max: 87.0%) and demonstrates strong industrial generalization on real-world semiconductor scanning electron microscope (SEM) production-line data.

Technology Category

• Computer Vision: Language and Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Anomaly Detection (AD) focuses on detecting samples that differ from the standard pattern, making it a vital tool in process control. Logical anomalies may appear visually normal yet violate predefined constraints on object presence, arrangement, or quantity, depending on reasoning and explainability. We introduce LogicQA, a framework that enhances AD by providing industrial operators with explanations for logical anomalies. LogicQA compiles automatically generated questions into a checklist and collects responses to identify violations of logical constraints. LogicQA is training-free, annotation-free, and operates in a few-shot setting. We achieve state-of-the-art (SOTA) Logical AD performance on public benchmarks, MVTec LOCO AD, with an AUROC of 87.6 percent and an F1-max of 87.0 percent along with the explanations of anomalies. Also, our approach has shown outstanding performance on semiconductor SEM corporate data, further validating its effectiveness in industrial applications.

Problem

• Detects logical anomalies violating object constraints.
• Provides explainable anomaly detection without training.
• Achieves SOTA performance on industrial benchmarks.

Innovation

• Training-free logical anomaly detection framework
• Automatically generates questions for anomaly checks
• Achieves SOTA performance without annotations