Normality Addition via Normality Detection in Industrial Image Anomaly Detection Models
This addresses the need for adaptable anomaly detection in industrial settings, but it is incremental as it builds on existing models and datasets.
The paper tackles the problem of updating anomaly detection models when definitions of normality change, by proposing a method called NAND that uses vision-language models to adjust decision boundaries post-training. They demonstrate its effectiveness on the MVTec AD benchmark dataset.
The task of image anomaly detection (IAD) aims to identify deviations from normality in image data. These anomalies are patterns that deviate significantly from what the IAD model has learned from the data during training. However, in real-world scenarios, the criteria for what constitutes normality often change, necessitating the reclassification of previously anomalous instances as normal. To address this challenge, we propose a new scenario termed "normality addition," involving the post-training adjustment of decision boundaries to incorporate new normalities. To address this challenge, we propose a method called Normality Addition via Normality Detection (NAND), leveraging a vision-language model. NAND performs normality detection which detect patterns related to the intended normality within images based on textual descriptions. We then modify the results of a pre-trained IAD model to implement this normality addition. Using the benchmark dataset in IAD, MVTec AD, we establish an evaluation protocol for the normality addition task and empirically demonstrate the effectiveness of the NAND method.