Mauro Barni

A training-free agentic framework built around specialist detectors, per-detector triage, and conflict-aware evidence arbitration is studied, indicating that the main challenge of open-world image forensics is not detecting manipulations, but calibrating trust in specialized forensic tools and arbitrating conflicting evidence.

Attribution accuracy peaks at intermediate representation levels, revealing that the right balance of representation and reference selection is crucial for effective synthetic image attribution.

Adversarial examples become significantly easier to generate as input dimensionality increases, challenging existing theoretical assumptions about their emergence.