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
I have the same question, at the same time, I am curious about the accuracy of depth estimation at a distance of more than 50m at the medium image resolution of 960x640. By the way, considering that the Adaptive Query Generation proposed in far3d actually contains the auxiliary task of depth estimation, how to determine whether the performance improvement in the ablation experiment is due to depth estimation as an auxiliary training task, or whether it actually comes from the query generated by the model? Looking forward to the author's reply and discussion.
Hello,
I noticed that the detection range for this work is 150 metres. Why is the maximum depth of the depth estimation network 110 metres?
The text was updated successfully, but these errors were encountered: