You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Generally speaking, our proposed framework could work under moving cameras since the original NeRF was designed for the multi-view setting.
But the performance might compromise when the input camera poses are not accurate enough or the input images contain many artifacts, like motion blur. Also, you may need to modify some training options, e.g., disabling NDC.
You put the camera in a fixed single view point, if my camera moves, does this code still work effectively?
The text was updated successfully, but these errors were encountered: