KP-SLAM: Joint Flow-Pointmap Prior Synchronization for Robust Consistent Dense Mapping
KP-SLAM is proposed, which predicts dense optical flow and paired pointmap priors from a shared representation and incorporates them into the same BA backend and introduces a Depth-Scale-Pose-to-Pointmap (DSPP) objective that relates optimized inverse depth, edge-wise relative scale, and camera pose to paired pointmap constraints.