Fredrik Ohlsson
Papers
1
Total Citations
8
H-Index
1
About
Fredrik Ohlsson is a researcher at the forefront of computer vision and geometric deep learning, with a particular focus on spherical and omnidirectional imaging. His most notable contribution is the development of HEAL-SWIN, a novel Vision Transformer architecture designed to operate directly on the sphere, eliminating the distortion and projection losses that plague conventional neural networks when processing wide-angle fisheye images. This work, published in 2024 and already garnering 8 citations, addresses a critical bottleneck in robotics applications such as autonomous driving, where high-resolution spherical data is increasingly vital. Ohlsson’s research bridges the gap between advanced geometric representations and practical deployment, offering a principled alternative to traditional planar image processing. By enabling efficient, distortion-free learning on the sphere, his work has the potential to significantly improve perception systems in real-world environments. His achievements underscore a commitment to solving fundamental challenges in representation learning, making him a rising voice in the intersection of computer vision and robotics.
Research Focus
Key Achievements
Top Papers
- 1HEAL-SWIN: A Vision Transformer on the Sphere8 citations · 2024