Author

A. E. Saddik

h-index 31 View on Semantic Scholar

We have 1 of 249 papers

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Preprint Jul 2026

Emergent Region-Level Facial Correspondence in Frozen Vision Foundation Models

These results establish frozen DINOv3 as a strong zero-shot representation for region-level facial correspondence and identify intermediate self-supervised features as the most useful layer for dense face analysis.

Izaldein Al-Zyoud, A. E. Saddik · 0 citations