Lecture 27: Inverse Graphics/Research In Angjoo's Group (19)
SainanChen
I guess prior assumptions about ping pong would definitely help with the accuracy of the model although it can be extrapolated from general human movement data.
han20192019
I think processing videos would be much slower than single photos since it needs much more computation. How can we deal with the problems with speed? Do we just use a more advanced devices or we have some different algorithms for video processing and image processing?
han20192019
I think processing videos would be much slower than single photos since it needs much more computation. How can we deal with the problems with speed? Do we just use a more advanced devices or we have some different algorithms for video processing and image processing?
bjmhs
Does the model need to be trained specifically with videos of athletes playing ping pong? If not, does training the model with ping pong videos gives a more accurate mesh recovery for ping pong video?
I guess prior assumptions about ping pong would definitely help with the accuracy of the model although it can be extrapolated from general human movement data.
I think processing videos would be much slower than single photos since it needs much more computation. How can we deal with the problems with speed? Do we just use a more advanced devices or we have some different algorithms for video processing and image processing?
I think processing videos would be much slower than single photos since it needs much more computation. How can we deal with the problems with speed? Do we just use a more advanced devices or we have some different algorithms for video processing and image processing?
Does the model need to be trained specifically with videos of athletes playing ping pong? If not, does training the model with ping pong videos gives a more accurate mesh recovery for ping pong video?