Federico Pernici

Federico Pernici serves as an Associate Professor at the University of Florence, specializing in computer vision and machine learning research.

His research program demonstrates a coherent progression from foundational multi-view geometry and 3D reconstruction to advanced incremental learning systems. Early work focused on incremental camera tracking (SLAM with zooming cameras) and lifelong object tracking in video streams, which naturally evolved into incremental appearance learning through deep neural networks. Current research centers on backward-compatible representations—learned features that maintain interchangeability across evolving model architectures without wasting training data, directly addressing catastrophic forgetting challenges in foundation model development. His overarching objective remains the creation of vision systems capable of incremental lifelong learning while preserving representation compatibility and prior knowledge retention.

No scientific awards were documented in the provided text.

No details regarding student advising, research grants, or funding sources were mentioned.

Information about laboratory facilities, research teams, or collaborative groups was not included in the source material.