search

ispc-lab / HRegNet

[ICCV 2021] HRegNet: A Hierarchical Network for Large-scale Outdoor LiDAR Point Cloud Registration
MIT License
90 stars 12 forks source link

Detect keypoints and find correspondances #7

Closed sYANGunique1111 closed 2 years ago

sYANGunique1111 commented 2 years ago

Hello,

I have been studying your model and I found out that the keypoints are found by summing the product of weights and knn points, how ever this may not affirm that they exactly exist in the original point cloud. Besides I found something silimar in the process of finding correspondances in target point cloud that you use the attentive weights to multiplys knn points as well, which doesn't verify the corres points exist in target point cloud either. Did I get somewhere wrong ?

FanLu97 commented 2 years ago

Hello, Your understanding is right. In LiDAR point cloud, a source point typically can not find the exactly accurate corresponding target point due to the sparsity. Thus, the detected keypoints would not perfectly coincide with the raw point cloud. Our detection aims to find salient virtual points. For finding correspondence, we search for K candidate corresponding points, which can weaken the negative effect caused by imperfect descriptors. For example, the corresponding point may be the second nearest neighboring in descriptor space. Thus, our design aims to assign higher weights to more confident candidate points.

FanLu97 commented 2 years ago

Close due to inactivity.