vx， vy in FCOS3D
#2125
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
{{ message }}
-
In the FCOS3D paper, the regression branch also predicts vx and vy (velocity?)
Just want to know why this information is helpful to 3D detection. No matter whether still or moving, a vehicle will have the same 3d bounding boxes if in the same pose.
Beta Was this translation helpful? Give feedback.
All reactions