Matt Uyttendaele

Papers

Total Citations

H-Index

About

Matt Uyttendaele is a leading researcher in computer vision and augmented reality, whose work has significantly advanced real-time 3D localization and scene understanding. His most influential contribution is a pioneering method for real-time monocular image-based 6-DoF (six-degree-of-freedom) localization, which enables continuous camera pose estimation from video by tracking natural features and matching them to offline structure-from-motion reconstructions. This work, cited over 50 times, has been foundational for applications in mobile augmented reality and autonomous navigation. Beyond this, Uyttendaele has made notable contributions to computational photography and image-based rendering, including high-dynamic-range imaging and video stabilization techniques that have been integrated into consumer products. His research bridges the gap between robust offline reconstruction and efficient online tracking, solving critical challenges in real-world AR deployment. With a career spanning both academic publications and industry impact at Microsoft Research, Uyttendaele’s work continues to influence how devices perceive and interact with their environments, making him a key figure in the evolution of practical computer vision systems.