Multi-scale sampling and feature fusion for dynamic human rendering

doi:10.1016/j.jvcir.2026.104766

Multi-scale sampling and feature fusion for dynamic human rendering

2026-03-05

0

PRE

AI

DOI：10.1016/j.jvcir.2026.104766

Original

Original request for help

Save

Abstract

En 中文

• MSFFNeRF, a novel framework that integrates conical volume sampling to enhance dynamic human rendering with improved detail fidelity. • Designs a multi-sampling and adaptive feature fusion module that effectively refines latent code features to improve appearance modeling of dynamic humans. • Achieves competitive performance gains over existing approaches on standard benchmarks.

Keywords:

MSFFNeRF

conical volume sampling

adaptive feature fusion

dynamic human rendering

latent code refinement

Multi-scale sampling and feature fusion for dynamic human rendering

Abstract

Journal

Organization