One-to-More: High-Fidelity Training-Free Anomaly Generation with Attention Control

Researchers have developed O2MAG, a training-free method for generating high-fidelity anomalies from a single anomalous image, which is crucial for improving industrial anomaly detection systems where real anomalous data are scarce. By leveraging self-attention and incorporating an anomaly mask to guide synthesis, O2MAG enhances the realism of generated anomalies, outperforming existing methods in downstream tasks.

Read the full article at arXiv cs.CV (Vision)

---

Want to create content about this topic? Use Nemati AI tools to generate articles, social posts, and more.