Pascal Frossard
Papers
6
Total Citations
217
H-Index
6
About
Pascal Frossard is a leading researcher in visual computing and machine learning, with a focus on omnidirectional imaging, graph-based signal processing, and human-robot interaction. His major contributions include pioneering graph-based classification methods for omnidirectional images, which address the geometric distortions inherent in 360° content, and developing geometry-aware convolutional filters that improve representation learning for spherical data. These works, cited over 140 times collectively, have advanced applications in robotics, virtual reality, and autonomous navigation. Frossard also contributed to benchmarking human-to-robot handovers, enabling robots to estimate physical properties of unseen objects in real time—a key step toward safer human-robot collaboration. More recently, his work on hierarchical training of deep neural networks using early exiting (2024) tackles resource-efficient AI for edge devices, reducing communication costs and privacy risks. With a strong track record of high-impact publications, Frossard’s research bridges theoretical innovation and practical deployment, making him a key figure in the evolution of immersive visual systems and intelligent robotics.
Research Focus
Key Achievements
Top Papers
- 1Visual Distortions in 360° Videos72 citations · 2019
- 2Graph-Based Classification of Omnidirectional Images69 citations · 2017
- 3
- 4Graph-Based Classification of Omnidirectional Images16 citations · 2017
- 5Hierarchical Training of Deep Neural Networks Using Early Exiting11 citations · 2024
- 6