notesum.ai

Published at November 15

GGAvatar: Reconstructing Garment-Separated 3D Gaussian Splatting Avatars from Monocular Video

cs.CV

cs.AI

cs.MM

Released Date: November 15, 2024

Authors: Jingxuan Chen¹

Aff.: ¹Jinan University-University of Birmingham Joint Institute, Guangzhou, Guangdong, China

Arxiv: http://arxiv.org/abs/2411.09952v1

Refer to caption

Methods (time)	male-3-casual	male-4-casual	female-3-casual	female-4-casual
	PSNR SSIM LPIPS	PSNR SSIM LPIPS	PSNR SSIM LPIPS	PSNR SSIM LPIPS
Neural Body( $>$ 0.5d)	24.94 0.9428 0.0326	24.71 0.9469 0.0423	23.87 0.9504 0.0346	24.37 0.9451 0.0382
InstantAvatar( $\sim$ 5m)	29.53 0.9716 0.0155	27.67 0.9626 0.0307	27.66 0.9709 0.0210	29.11 0.9683 0.0167
GaussianAvatar( $\sim$ 0.5h)	30.98 0.9790 0.0145	28.78 0.9755 0.0228	29.55 0.9762 0.0225	30.84 0.9771 0.0140
GART( $\sim$ 2m)	30.73 0.9769 0.0360	27.58 0.9679 0.0602	26.60 0.9659 0.0466	29.23 0.9721 0.0378
SCARF( $>$ 2d)†	30.59 0.9770 0.0247	28.99 0.9701 0.0257	30.14 0.9764 0.0282	29.96 0.9718 0.0267
SCARF( $\sim$ 20m)†	18.20 0.8521 0.3578	18.41 0.8860 0.3618	19.90 0.9120 0.2494	19.76 0.8920 0.2274
GGAvatar( $\sim$ 20m)†	31.01 0.9812 0.0349	29.40 0.9728 0.0509	27.96 0.9753 0.0452	30.29 0.9750 0.0343