Multi-View Industrial Anomaly Detection with Epipolar Constrained Cross-View Fusion
This work addresses anomaly detection in industrial settings using multi-camera systems, offering an incremental improvement over existing multi-view methods.
The paper tackled the problem of multi-view industrial anomaly detection by introducing an epipolar geometry-constrained attention module and a pretraining strategy, resulting in outperforming existing methods on a state-of-the-art dataset.
Multi-camera systems provide richer contextual information for industrial anomaly detection. However, traditional methods process each view independently, disregarding the complementary information across viewpoints. Existing multi-view anomaly detection approaches typically employ data-driven cross-view attention for feature fusion but fail to leverage the unique geometric properties of multi-camera setups. In this work, we introduce an epipolar geometry-constrained attention module to guide cross-view fusion, ensuring more effective information aggregation. To further enhance the potential of cross-view attention, we propose a pretraining strategy inspired by memory bank-based anomaly detection. This approach encourages normal feature representations to form multiple local clusters and incorporate multi-view aware negative sample synthesis to regularize pretraining. We demonstrate that our epipolar guided multi-view anomaly detection framework outperforms existing methods on the state-of-the-art multi-view anomaly detection dataset.