You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
maybe I did not search enough, but I can not find the header information of the .txt files where the instance segmentation point clouds are released. Once downloaded the dataset, I checked that all .txt files has 8 columns, and I already know the meaning of the 6 firsts (x,y,z,r,g,b), but the last two are unknown for me.
As I could saw, taking the '1_points_GTv3.txt' file as an example, 7th column has only values between 0 and 14, which makes sense to think that this is the label column. However, 8th column has 1655 values in the range of [0,2736] and also -100. In the documentation related to instance segmentation is written that all ground points are labelled with -100 values, but this value appears only in the 8th column and not in the 7th. This, together with the fact that 8th column has values in the range mentioned previously, tricks me.
Sorry for opening this issue, but this should be clearer
The text was updated successfully, but these errors were encountered:
maybe I did not search enough, but I can not find the header information of the .txt files where the instance segmentation point clouds are released. Once downloaded the dataset, I checked that all .txt files has 8 columns, and I already know the meaning of the 6 firsts (x,y,z,r,g,b), but the last two are unknown for me.
As I could saw, taking the '1_points_GTv3.txt' file as an example, 7th column has only values between 0 and 14, which makes sense to think that this is the label column. However, 8th column has 1655 values in the range of [0,2736] and also -100. In the documentation related to instance segmentation is written that all ground points are labelled with -100 values, but this value appears only in the 8th column and not in the 7th. This, together with the fact that 8th column has values in the range mentioned previously, tricks me.
Sorry for opening this issue, but this should be clearer
Hi! Maybe I can help , here is the format for txt files :x,y,z,r,g,b,semantic label,instance label.So the numbers range [0,14] at 7th column means here are 15 classes in the dataset; numbers range [0,2736] at 8th column means instance labels , but it may not consistent from 0 to 2736.As for lable -100 it means which will be ignored in the model . It's better to see the code in HAIS/data/prepare_data_inst_instance_stpls3d.py, it may help you to understand the corresponding meaning better.
Hi,
maybe I did not search enough, but I can not find the header information of the .txt files where the instance segmentation point clouds are released. Once downloaded the dataset, I checked that all .txt files has 8 columns, and I already know the meaning of the 6 firsts (x,y,z,r,g,b), but the last two are unknown for me.
As I could saw, taking the '1_points_GTv3.txt' file as an example, 7th column has only values between 0 and 14, which makes sense to think that this is the label column. However, 8th column has 1655 values in the range of [0,2736] and also -100. In the documentation related to instance segmentation is written that all ground points are labelled with -100 values, but this value appears only in the 8th column and not in the 7th. This, together with the fact that 8th column has values in the range mentioned previously, tricks me.
Sorry for opening this issue, but this should be clearer
The text was updated successfully, but these errors were encountered: