How to get the segmented points？

[slowlypasser]

I run the fpcc_test.py，and get the results.I only see the center score prediction，but i want to get the instance prediction.Can you tell me how to get the instance predition?Thanks!

[xyjbaal]

The code of output instance prediction is written in L294 of fpcc_test.py. cloud you check this part again?

[slowlypasser]

The code of output instance prediction is written in L294 of fpcc_test.py. cloud you check this part again?

I have seen that,but i use open3d to visualize all the results(all the ".txt" files).The point clouds are same like this.

[xyjbaal]

Could you tell me the name of .txt file? xxxx_grouppred.txt or xxxx_pred.txt?

[slowlypasser]

Could you tell me the name of .txt file? xxxx_grouppred.txt or xxxx_pred.txt? Yes,xxxx_grouppred.txt and xxxx_pred.txt is same.So does the test file. Like 0000_026.txt 0000_026_grouppred.txt 0000_026_pred.txt,they are the same in visualization.

[slowlypasser]

Could you tell me the name of .txt file? xxxx_grouppred.txt or xxxx_pred.txt? Yes,xxxx_grouppred.txt and xxxx_pred.txt is same.So does the test file. Like 0000_026.txt 0000_026_grouppred.txt 0000_026_pred.txt,they are the same in visualization.

I want to get the result like the paper,and then extract the instance one by one.

[xyjbaal]

I suppose there are 6 columns in _grouppred.txt. the first three columns are the coordinates of the point and the last three columns should be RGB.

for _pred.txt, the last column is the predicted instance label.

[slowlypasser]

I suppose there are 6 columns in _grouppred.txt. the first three columns are the coordinates of the point and the last three columns should be RGB.

for _pred.txt, the last column is the predicted instance label.

Ok,i get it.Maybe,it's the problem of visualization.I will try to visualize it using aother ways.

[xyjbaal]

That's great. Maybe you can try using CloudCompare. it is a free software.