search

fabiopoiesi / dip

Project page of the paper "Distinctive 3D local deep descriptors" accepted in IEEE International Conference on Pattern Recognition 2020.
65 stars 4 forks source link

How to find the overlap region？ #11

Closed wangzz15 closed 2 years ago

wangzz15 commented 3 years ago

Dear author, thanks for your excellent work. I'm a beginner in the filed of Point Cloud Registration. I still don't know how to find the overlap region. The red points seems to be randomly selected in the source code. Could you please give some advice? image

fabiopoiesi commented 3 years ago

The overlap region can be found using knn, setting k=1, on the point clouds transformed with their respective transformations. Both the point clouds should be in the same reference system to apply knn. You should also apply a distance threshold on the nn point.

In the supplementary material of SmoothNet paper, this process is explained in detail. In DIP I used the data pre-processed in this paper. However, it is something that can be code up in minutes.

The red points are indeed randomly sampled, using farthest point sampling, from the cyan points.

wenli4313 commented 2 years ago

I have been studying the DIP paper recently, and I would like to ask you some questions. If you see the message, you can contact me on wechat:zwl577729391, thank you very much.