CHRIS: Clothed Human Reconstruction with Side View Consistency

TL;DR

CHRIS introduces a novel method for clothed human reconstruction from a single RGB image by incorporating side view consistency, significantly improving realism and surface accuracy.

Contribution

The paper proposes a Side-View Normal Discriminator and Multi-to-One Gradient Computation to enhance global and local consistency in human reconstruction from single images.

Findings

Achieves state-of-the-art results on public benchmarks.

Outperforms previous methods in realism and surface accuracy.

Effectively utilizes side-view information despite single-view input.

Abstract

Creating a realistic clothed human from a single-view RGB image is crucial for applications like mixed reality and filmmaking. Despite some progress in recent years, mainstream methods often fail to fully utilize side-view information, as the input single-view image contains front-view information only. This leads to globally unrealistic topology and local surface inconsistency in side views. To address these, we introduce Clothed Human Reconstruction with Side View Consistency, namely CHRIS, which consists of 1) A Side-View Normal Discriminator that enhances global visual reasonability by distinguishing the generated side-view normals from the ground truth ones; 2) A Multi-to-One Gradient Computation (M2O) that ensures local surface consistency. M2O calculates the gradient of a sampling point by integrating the gradients of the nearby points, effectively acting as a smooth operation. Experimental results demonstrate that CHRIS achieves state-of-the-art performance on public benchmarks and outperforms the prior work.