notesum.ai

Published at November 29

Fleximo: Towards Flexible Text-to-Human Motion Video Generation

cs.CV

cs.LG

Released Date: November 29, 2024

Authors: Yuhang Zhang¹, Yuan Zhou², Zeyu Liu³, Yuxuan Cai², Qiuyue Wang², Aidong Men¹, Huan Yang²

Aff.: ¹Beijing University of Posts and Telecommunications; ²01.AI; ³Tsinghua University

Arxiv: http://arxiv.org/pdf/2411.19459v1

Refer to caption

Methods	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$	DreamSim $\uparrow$	FID $\downarrow$	FVD $\downarrow$	MotionScore $\uparrow$
I2VGen-XL [42]	7.931	0.3684	0.559	0.537	220.221	1905.26	0.6806
VideoCrafter [3]	6.727	0.5190	0.608	0.211	149.534	1536.00	0.6866
DynamiCrafter [35]	9.607	0.6800	0.407	0.699	99.644	1462.42	0.6868
Fleximo	16.647	0.7148	0.284	0.879	76.181	1360.12	0.6990