CrowdGaussian：从单张图像重建高保真3D高斯人体人群
CrowdGaussian: Reconstructing High-Fidelity 3D Gaussians for Human Crowd from a Single Image
提出自监督适应流程，使预训练人体模型能从严重遮挡的单张图像中重建完整3D人体，并利用自校准学习策略微调单步扩散模型以提升多人3DGS表示质量，可直接用于拥挤场景的三维重建与渲染。
原文依据 we devise a self-supervised adaptation pipeline that enables the pretrained large human model to reconstruct complete 3D humans
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论单视图三维人体重建近年来受到广泛关注。尽管已有诸多进展，但先前研究集中于从清晰、近距离的单人图像重建三维模型，在更普遍的多人物场景中往往效果欠佳。重建三维人体人群模型是一项高度复杂的任务，面临诸多挑战：1）大量遮挡，2）清晰度低，3）外观数量多且多样。为解决该任务，我们提出CrowdGaussian，一个直接从单张图像输入重建多人三维高斯泼溅（3DGS）表示的统一框架。为处理遮挡，我们设计了一种自监督适应流程，使预训练大型人体模型能够从严重遮挡的输入中重建具有合理几何与外观的完整三维人体。此外，我们引入自校准学习（SCL）。该训练策略通过混合身份保持样本与干净/损坏图像对，使单步扩散模型能够自适应地将粗糙渲染细化为最佳质量。输出可被蒸馏回以增强多人3DGS表示的质量。大量实验表明，CrowdGaussian能生成多人场景的逼真、几何一致的重建结果。
英文原文摘要 1369 字符
Single-view 3D human reconstruction has garnered significant attention in recent years. Despite numerous advancements, prior research has concentrated on reconstructing 3D models from clear, close-up images of individual subjects, often yielding subpar results in the more prevalent multi-person scenarios. Reconstructing 3D human crowd models is a highly intricate task, laden with challenges such as: 1) extensive occlusions, 2) low clarity, and 3) numerous and various appearances. To address this task, we propose CrowdGaussian, a unified framework that directly reconstructs multi-person 3D Gaussian Splatting (3DGS) representations from single-image inputs. To handle occlusions, we devise a self-supervised adaptation pipeline that enables the pretrained large human model to reconstruct complete 3D humans with plausible geometry and appearance from heavily occluded inputs. Furthermore, we introduce Self-Calibrated Learning (SCL). This training strategy enables single-step diffusion models to adaptively refine coarse renderings to optimal quality by blending identity-preserving samples with clean/corrupted image pairs. The outputs can be distilled back to enhance the quality of multi-person 3DGS representations. Extensive experiments demonstrate that CrowdGaussian generates photorealistic, geometrically coherent reconstructions of multi-person scenes.