Search papers, labs, and topics across Lattice.
The paper introduces VID-AD, a new dataset for image-level logical anomaly detection in industrial settings, designed to evaluate robustness against vision-induced distractions like background clutter and illumination shifts. VID-AD comprises 10 manufacturing scenarios with controlled variations in visual appearance across 5 capture conditions, resulting in 50 one-class anomaly detection tasks. The authors also propose a language-based anomaly detection framework that uses contrastive learning on text descriptions generated from normal images to focus on logical attributes rather than low-level visual features, demonstrating improved performance over baselines.
Current anomaly detection benchmarks are too easily fooled by visual noise; VID-AD offers a rigorous new testbed to isolate and address this weakness.
Logical anomaly detection in industrial inspection remains challenging due to variations in visual appearance (e.g., background clutter, illumination shift, and blur), which often distract vision-centric detectors from identifying rule-level violations. However, existing benchmarks rarely provide controlled settings where logical states are fixed while such nuisance factors vary. To address this gap, we introduce VID-AD, a dataset for logical anomaly detection under vision-induced distraction. It comprises 10 manufacturing scenarios and five capture conditions, totaling 50 one-class tasks and 10,395 images. Each scenario is defined by two logical constraints selected from quantity, length, type, placement, and relation, with anomalies including both single-constraint and combined violations. We further propose a language-based anomaly detection framework that relies solely on text descriptions generated from normal images. Using contrastive learning with positive texts and contradiction-based negative texts synthesized from these descriptions, our method learns embeddings that capture logical attributes rather than low-level features. Extensive experiments demonstrate consistent improvements over baselines across the evaluated settings. The dataset is available at: https://github.com/nkthiroto/VID-AD.
