Inference data

[scw0819]

Hi, thank you for your contribution of code, which can solve many problems and achieve better results in panoramic segmentation of large-scale lidar point clouds. I noticed that in the data in the NPM3D dataset, it is [N,6], which not only includes the value of xyz, but also 'scalar_reflectance', 'scalar_class', 'scalar_label', and the same is true for the test set. But when I want to reason about my own data, my data format is [N,4] and there are no semantic categories and object instances. How should I reason about my own data, or where should I get the semantic categories and object instances of my own data? I didn't solve this problem in the code, thank you for your generous answer.

[bxiang233]

Hi, thank you for your contribution of code, which can solve many problems and achieve better results in panoramic segmentation of large-scale lidar point clouds. I noticed that in the data in the NPM3D dataset, it is [N,6], which not only includes the value of xyz, but also 'scalar_reflectance', 'scalar_class', 'scalar_label', and the same is true for the test set. But when I want to reason about my own data, my data format is [N,4] and there are no semantic categories and object instances. How should I reason about my own data, or where should I get the semantic categories and object instances of my own data? I didn't solve this problem in the code, thank you for your generous answer.

Hi, thanks a lot for youe interest in my work! sorry for the late reply, I just finished a ddl. I think I answered the same question before, for your reference: https://github.com/bxiang233/PanopticSegForLargeScalePointCloud/issues/2#issuecomment-1655269126

[scw0819]

Thank you for your answer. In fact, I have already done so and achieved certain results. Thank you for your help.

[bxiang233]

Thank you for your answer. In fact, I have already done so and achieved certain results. Thank you for your help.

Good to hear! you are welcome.