Riccardo Spezialetti

Papers

Total Citations

H-Index

About

Riccardo Spezialetti is a computer vision researcher whose work bridges 3D scene understanding, neural rendering, and geometric deep learning. His most prominent contribution is the development of ReNe (Relighting NeRF), a richly annotated dataset designed to advance novel view synthesis and relighting of real-world objects using Neural Radiance Fields (NeRF). By capturing objects under one-light-at-a-time (OLAT) conditions, this dataset addresses one of the most challenging frontiers in photorealistic rendering — synthesizing scenes under previously unobserved lighting — and has already garnered over 40 citations since its 2023 publication, signaling strong community uptake. Beyond neural rendering, Spezialetti has made meaningful contributions to 3D surface analysis, notably through his 2020 work on self-supervised spherical CNNs for learning canonical surface orientations. Rather than relying on handcrafted geometric heuristics, this approach leverages learned representations to determine consistent surface normals, with broad implications for robotics and 3D shape understanding. Taken together, his research reflects a commitment to building principled, data-driven solutions for complex geometric and photometric challenges, making him a noteworthy emerging voice in the 3D computer vision community.