search

zhanglonghao1992 / One-Shot_Free-View_Neural_Talking_Head_Synthesis

Pytorch implementation of paper "One-Shot Free-View Neural Talking-Head Synthesis for Video Conferencing"
Other
735 stars 141 forks source link

some questions about dense_motion.py #45

Open Winfredy opened 2 years ago

Winfredy commented 2 years ago

delete detect

according to the author's method, we need to multipy coordinate_grid and Rs*inv(Rd). But why you use coordinate_grid directly?