search

octree-nn / octformer

OctFormer: Octree-based Transformers for 3D Point Clouds
MIT License
259 stars 18 forks source link

ScanNet Segmentation #2

Closed fyj534609 closed 1 year ago

fyj534609 commented 1 year ago

Hello, I would like to ask whether when downloading the ScanNet Segmentation dataset, do I need to download the complete dataset of 1.3TB, or is it possible to download only a part of it?

wang-ps commented 1 year ago

I downloaded the whole dataset the first time I tried to process ScanNet. The dataset after processing is small and is only about 15G.

fyj534609 commented 1 year ago

Thank you for your response. So should I download the whole 1.3TB dataset and then use the command 'python tools/seg_scannet.py --run process_scannet --path_in ' to process the downloaded dataset, resulting in about 15GB of remaining data?

wang-ps commented 1 year ago

Maybe, you can choose not to download the RGBD scan frames. An example is shown below. The following is all you need to generate the required dataset

image

fyj534609 commented 1 year ago

Maybe, you can choose not to download the RGBD scan frames. An example is shown below. The following is all you need to generate the required dataset

image

Thank you for your response. Could you please confirm if the final downloaded file directory should look like this? ├── scannet │ ├── scans │ │ ├── [scene_id] │ │ │ ├── [scene_id].aggregation.json │ │ │ ├── [scene_id].txt │ │ │ ├── [scene_id]_vh_clean.aggregation.json │ │ │ ├── [scene_id]_vh_clean.segs.json │ │ │ ├── [scene_id]_vh_clean_2.0.010000.segs.json │ │ │ ├── [scene_id]_vh_clean_2.labels.ply │ │ │ ├── [scene_id]_vh_clean_2.ply │ ├── scans_test │ │ ├── [scene_id] │ │ │ ├── [scene_id].aggregation.json │ │ │ ├── [scene_id].txt │ │ │ ├── [scene_id]_vh_clean.aggregation.json │ │ │ ├── [scene_id]_vh_clean.segs.json │ │ │ ├── [scene_id]_vh_clean_2.0.010000.segs.json │ │ │ ├── [scene_id]_vh_clean_2.ply │ ├── scannetv2-labels.combined.tsv 1691115392320

wang-ps commented 1 year ago

Yes, very informative and clear figure!

wang-ps commented 1 year ago

I downloaded the whole dataset the first time I tried to process ScanNet. The dataset after processing is small and is only about 15G.

Correct: the dataset after process is about 8G

fyj534609 commented 1 year ago

Yes, very informative and clear figure!

Thank you very much for your reply!