Xuxiang Feng

Papers

Total Citations

H-Index

About

Xuxiang Feng is a researcher specializing in computer vision and 3D scene understanding, with a particular focus on neural radiance fields (NeRF) and their application to 6D pose estimation. Their most notable contribution is the development of C2Fi-NeRF, a coarse-to-fine inversion framework that leverages NeRF’s implicit scene representation to achieve robust and accurate 6D object pose estimation from a single image. This work addresses a critical challenge in robotics and augmented reality: enabling precise object localization without requiring extensive 3D models or dense viewpoint coverage. By introducing a hierarchical inversion strategy, Feng’s method improves both efficiency and accuracy, offering a practical solution for real-world applications. While still early in its impact, with 3 citations since its 2024 publication, the paper represents a promising direction in the fusion of neural rendering and pose estimation. Feng’s research sits at the intersection of implicit neural representations and geometric computer vision, contributing to the growing body of work that seeks to make NeRF-based techniques more accessible for downstream tasks. Their work is particularly relevant for researchers exploring end-to-end learning for spatial reasoning and scene interaction.