search

CMU-Perceptual-Computing-Lab / MonocularTotalCapture

Code for CVPR19 paper "Monocular Total Capture: Posing Face, Body and Hands in the Wild"
http://domedb.perception.cs.cmu.edu/mtc.html
662 stars 123 forks source link

About time consuming? #36

Open Arthur151 opened 4 years ago

Arthur151 commented 4 years ago

Does anyone knows the per-image time consuming of this method? Thinks!

xiangdonglai commented 4 years ago

The whole pipeline takes a lot of steps. OpenPose -> POF prediction -> single frame fitting -> tracking. The current implementation process all frames in one stage before starting the next stage. Roughly estimate that on average it will take 30 seconds every frame.