Gordon Wetzstein
Papers
12
Total Citations
1,247
H-Index
10
About
Gordon Wetzstein is a pioneering computational imaging researcher whose work has fundamentally advanced how machines perceive the world beyond direct line of sight. Best known for developing breakthrough non-line-of-sight (NLOS) imaging algorithms, Wetzstein and his collaborators introduced the light-cone transform for confocal NLOS imaging (492 citations) and the fast f-k migration framework (257 citations), enabling cameras to effectively "see around corners" — a capability with profound implications for autonomous vehicles, robotics, and remote sensing. His 2020 work on confocal diffuse tomography (124 citations) extended these principles to imaging through scattering media like fog and rain, addressing critical challenges for real-world perception systems. Demonstrating remarkable breadth, Wetzstein translated NLOS concepts into the acoustic domain (102 citations) and explored end-to-end "dirty pixels" approaches that jointly optimize image acquisition and deep learning pipelines for raw sensor data. His investigations span time-of-flight depth sensing, light field photography, and most recently vision-language-action models for robotics. With thousands of citations accumulated across a decade of influential work, Wetzstein has established himself as one of the defining voices in computational photography and machine perception research.
Research Focus
Key Achievements
Top Papers
- 1Confocal non-line-of-sight imaging based on the light-cone transform492 citations · 2018
- 2Wave-based non-line-of-sight imaging using fast <i>f-k</i> migration257 citations · 2019
- 3
- 4Acoustic Non-Line-Of-Sight Imaging102 citations · 2019
- 5
- 6Dirty Pixels: Towards End-to-end Image Processing and Perception59 citations · 2021
- 7Computational imaging with multi-camera time-of-flight systems59 citations · 2016
- 8
- 9CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models23 citations · 2025
- 10