Neural Human Performer: Learning Generalizable Radiance Fields for Human Performance Rendering

TL;DR

This paper introduces Neural Human Performer, a method that learns generalizable neural radiance fields for synthesizing free-viewpoint videos of arbitrary human performances, effectively handling occlusions and articulations through temporal and multi-view transformers.

Contribution

It proposes a novel approach combining temporal and multi-view transformers with a parametric human body model for robust, generalizable human performance rendering.

Findings

Outperforms recent generalizable NeRF methods on unseen identities and poses.

Effectively handles occlusions and dynamic articulations.

Demonstrates superior performance on ZJU-MoCap and AIST datasets.

Abstract

In this paper, we aim at synthesizing a free-viewpoint video of an arbitrary human performance using sparse multi-view cameras. Recently, several works have addressed this problem by learning person-specific neural radiance fields (NeRF) to capture the appearance of a particular human. In parallel, some work proposed to use pixel-aligned features to generalize radiance fields to arbitrary new scenes and objects. Adopting such generalization approaches to humans, however, is highly challenging due to the heavy occlusions and dynamic articulations of body parts. To tackle this, we propose Neural Human Performer, a novel approach that learns generalizable neural radiance fields based on a parametric human body model for robust performance capture. Specifically, we first introduce a temporal transformer that aggregates tracked visual features based on the skeletal body motion over time. Moreover, a multi-view transformer is proposed to perform cross-attention between the temporally-fused features and the pixel-aligned features at each time step to integrate observations on the fly from multiple views. Experiments on the ZJU-MoCap and AIST datasets show that our method significantly outperforms recent generalizable NeRF methods on unseen identities and poses. The video results and code are available at https://youngjoongunc.github.io/nhp.