Luigi Di Stefano
Papers
18
Total Citations
186
H-Index
9
About
Luigi Di Stefano is a prominent computer vision and robotics researcher whose work spans stereo vision, 3D perception, neural rendering, and robotic systems. Based at the University of Bologna, Di Stefano has built a distinguished career bridging fundamental visual computing research with practical robotic applications. His contributions touch several interconnected domains. In neural rendering, his 2023 ReNe dataset for relighting Neural Radiance Fields has already garnered 37 citations, positioning him at the frontier of novel view synthesis research. His long-standing interest in stereo vision is reflected in rigorous benchmarking work evaluating efficient, real-time stereo algorithms suitable for resource-constrained robotic platforms. In 3D scene understanding, he has advanced both semantic segmentation — notably addressing domain-shift challenges by fusing 2D and 3D networks — and real-time mapping through the SkiMap framework, enabling robots to maintain efficient multi-resolution environmental representations. Di Stefano has also explored haptic interfaces for visually impaired users, GPU-accelerated 3D local descriptors, and self-supervised learning for surface orientation, demonstrating remarkable breadth. His work on deformable object segmentation and robotic grasping via multi-view reconstruction further underscores his commitment to grounding computer vision in real-world robotic utility. Across these contributions, his research consistently prioritizes both theoretical rigor and deployable, efficient solutions.
Research Focus
Key Achievements
Top Papers
- 1
- 2Perception of depth information by means of a wire-actuated haptic interface25 citations · 2002
- 3
- 4
- 5
- 6
- 7SkiMap++: Real-Time Mapping and Object Recognition for Robotics10 citations · 2017
- 8GPU-SHOT: Parallel Optimization for Real-Time 3D Local Description10 citations · 2013
- 9Learning to Orient Surfaces by Self-supervised Spherical CNNs9 citations · 2020
- 10SkiMap: An efficient mapping framework for robot navigation8 citations · 2017