ganlumomo / bkisemanticmapping Goto Github PK

Hi,

First of all, thank you for sharing such a great work! I was wondering if there was a specific code or method you used to calculate the area under curve for the map density? I would like to compare and benchmark a few methods and if it is possible, use that as one of the metrics. Could you let me know how it was done in your paper? Thank you

Hi, @ganlumomo

How are the quantitative results of the SemanticKITTI dataset obtained?

I used the semantickitti_04 data and evaluation code in this repository.
I modified config/semantickitti.yml to use all scans:

scan_num: 271

The results of executing the evaluation code are as follows:

[ 0.  1.  4.  5.  6.  9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.]
[0.00000000e+00 4.33485335e-03 0.00000000e+00 0.00000000e+00
 0.00000000e+00 7.63473164e-04 0.00000000e+00 1.14741928e-06
 8.34710744e-03 1.71361104e-06 9.96031937e-06 1.43424767e-04
 9.64897045e-06 1.47085490e-03 0.00000000e+00 0.00000000e+00]

Looking at the estimated semantic map, I see that many points are classified as unlabeled (class 0).
I suspect the parameter settings, but I don't know how to adjust them.

Do you know what's going on?

Hi Lu,

Hope you are doing well recently! Some students from the Mobile Robotics course did try to build this repo on Ubuntu 20 which is not initially supported. Is it possible to create a new branch adding these modifications? I also noticed the student already created a pull request. So you can directly see the changes.