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
Hey guys, thanks for your brilliant work in both MapTR and LaneGAP. After reading this paper, it seems that the perception pipeline for LaneGAP is basically the same with MapTR? And if not, what's the difference? Hope to get your reply soon
The text was updated successfully, but these errors were encountered:
LaneGAP and MapTR do share similar network design. LaneGAP extends MapTR to lane graph modeling. We integrade them into one unified framework in MapTRv2 (https://arxiv.org/abs/2308.05736).
The path-wise modeling in LaneGAP enables MapTR percepting lane graph with better performance (superior than pixel-wise, piece_wise) at least cost (without need of post-processing in pixel-wise and extra connectivity prediction in piece-wise).
Hey guys, thanks for your brilliant work in both MapTR and LaneGAP. After reading this paper, it seems that the perception pipeline for LaneGAP is basically the same with MapTR? And if not, what's the difference? Hope to get your reply soon
The text was updated successfully, but these errors were encountered: