huanghuang113

> I have written a visualization code in test_partseg.py, output point data to text file and see it(3d). If I submit the code to GitHub, is there still maintenance ongoing?...

> I'm trying to run SoftGroup on my custom dataset. I'm trying to detect leaves, nodes & stem on a plant. I also used some code of HIAS to generate...

how do you deal with this problem？

> Hi, I have successfully run my own dataset on mask3d using the ckpt. To check which offical dataset your data most similar to. Then follow the guide to preprocess...

> > 你好，幸运的是我也在研究类似于 stpls3d 的数据集。加工它并不是很难。首先，您应该使用 cloudcompare 或其他点云路径形式将数据子样本间隔为 0.3。然后将数据保存到 .txt 并确保每一行如下所示：x,y,z,r,g,b(x,y,z,r,g,b,sem_label, ins_label for train) 用“,”分割 。否则，您需要更改[Line120](https://github.com/JonasSchult/Mask3D/blob/b532e36212755c3b64b8888865ad550de0887fe4/datasets/preprocessing/stpls3d_preprocessing.py#L120C7-L120C59)。 如果您不想自己训练，Stpls3d 测试检查点在室外数据集中表现得非常好。你可以尝试一下。像这样输出。 顺便说一句，如果你有一个很大的场景数据集，你可以将其整体放置，而无需手动剪切。50 是分割数据集时预处理代码中的默认大小，如果您想加快处理速度，可以将其更改得更大。 > > 结果像这样  Thank you very much for such a...

> Hi, I am working to test on 'Synthetic_v3_InstanceSegmentation' test dataset. However, no matter how to split data to small blocks, it awalys come out "Segmentation fault (core dumped)" without...

> Hi, you need to add label for val not for test. The easiest way is to add 2 columns, like this x,y,z,r,g,b,100, 100. I apologize for my rude request,...