InCLAD: A Continual Learning Benchmark for Industrial Visual Anomaly Detection
Industrial anomaly detection benchmarks have advanced rapidly, yet they remain largely static and therefore underrepresent deployment settings in which inspection systems must be extended from one industrial component to another over time.
ProofPaper ↗
Key points
- In our setting, continual industrial anomaly detection is framed as a task-incremental problem: each component defines a task, while the learning objective remains semi-supervised.
- In addition to single-dataset continual scenarios, we propose challenging scenarios spanning across multiple datasets.
- We further evaluate continual performance from two complementary perspectives: image-level anomaly detection and pixel-level anomaly localization.
- By recasting industrial anomaly detection as a continual benchmark rather than a static train/test problem, our work establishes a concrete path toward studying knowledge retention, adaptation, and generalization in evolving industrial environments.
Sources (1)
- [1]InCLAD: A Continual Learning Benchmark for Industrial Visual Anomaly DetectionarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 10, 04:18 PM
Industrial anomaly detection benchmarks have advanced rapidly, yet they remain largely static and therefore underrepresent deployment settings in which inspection systems must be extended from one industrial component to another over time.
In our setting, continual industrial anomaly detection is framed as a task-incremental problem: each component defines a task, while the learning objective remains semi-supervised.
Extractive summary: sentences quoted from the sources.