Realistic Speech-Driven Talking Video Generation with Personalized Pose

Table 1

Mean Opinion Score (MOS) of 100 participants on 4 questions. Q1: completeness of body. Q2: the face is clear. Q3: the body movement is correlated with audio. Q4:overall quality.

	Q1	Q2	Q3	Q4
Learning gesture [31]	3.414	3.659	3.914	3.308
Neural-voice-puppetry[32]	3.202	3.840	3.180	3.542
EverybodyDance [33]	3.944	3.662	3.680	3.681
Personalized-bodyPose[29]	3.894	4.011	3.383	3.762
Our method	3.901	4.083	3.526	3.778