search

drprojects / superpoint_transformer

Official PyTorch implementation of Superpoint Transformer introduced in [ICCV'23] "Efficient 3D Semantic Segmentation with Superpoint Transformer" and SuperCluster introduced in [3DV'24 Oral] "Scalable 3D Panoptic Segmentation As Superpoint Graph Clustering"
MIT License
508 stars 65 forks source link

xy_tiling setting when process own data #103

Closed QingXia1994 closed 2 months ago

QingXia1994 commented 2 months ago

where should i modify when i need to SampleXYTiling by specific range like 100m*100m instead of by regular XY grid with xy_tiling².

thanks a lot !

drprojects commented 2 months ago

This is not a functionality we support. We only support SampleXYTiling and SampleRecursiveMainXYAxisTiling. These operate splitting operations "blindly" in the sense that they are not targeting a specific tile size but rather obey a required number of splits.

I agree this could be an interesting feature. This could be implemented within SampleXYTiling, as an extra argument. This would also mean that the following functions should be adapted: BaseDataset.get_tile_from_path(), BaseDataset.all_cloud_ids(), and BaseDataset._process_single_cloud().

Feel free to create a Pull Request if you implement this ! :wink: