Search papers, labs, and topics across Lattice.
1
0
2
4
VFMs like DINOv2 can now beat vision-language models at zero-shot anomaly detection, thanks to a new training scheme that overcomes dataset limitations and adapts VFMs more effectively.
