Marwan Taher

Papers

Total Citations

H-Index

About

Marwan Taher is a researcher at the forefront of 3D computer vision and robotics, with a focus on bridging the gap between efficient neural scene representations and real-world object interaction. His key research areas include 3D object pose estimation, neural rendering, and model-based robotic perception. Taher’s most notable contribution is his work on “Fit-NGP,” a method that leverages Neural Graphics Primitives (NGP) to achieve highly accurate and robust 3D object pose estimation from a single density field. By demonstrating that the efficient radiance field reconstructions from state-of-the-art methods are directly suitable for pose estimation, Taher has opened new pathways for deploying neural representations in challenging robotic applications, such as grasping and manipulation. Though early in his career, his work has already garnered attention for its practical impact, with Fit-NGP receiving its first citation in 2024. Taher’s research is distinguished by its focus on making advanced neural techniques computationally efficient and deployable in real-world settings, promising to accelerate the adoption of neural graphics primitives in autonomous systems and interactive robotics.