search

HpWang-whu / YOHO

[ACM MM 2022] You Only Hypothesize Once: Point Cloud Registration with Rotation-equivariant Descriptors
https://hpwang-whu.github.io/YOHO
134 stars 19 forks source link

thanks for your excellent work #7

Closed SingularityDog closed 1 year ago

SingularityDog commented 1 year ago

thank you so much,but how to fetch keypoints ?

HpWang-whu commented 1 year ago

Hi @SingularityDog , Thanks for your interest! We use 5k randomly-sampled points as keypoints on each scan. The keypoints are the same as that used in FCGF/PerfectMatch/SpinNet.

Sincerely,

SingularityDog commented 1 year ago

Is that sampled like farthest point sampling? or completely a random process?

SingularityDog commented 1 year ago

already found the answer in another issue,thanks a lot