search

Julie-tang00 / Point-BERT

[CVPR 2022] Pre-Training 3D Point Cloud Transformers with Masked Point Modeling
MIT License
541 stars 65 forks source link

Details about DGCNN #42

Open ham-stan opened 2 years ago

ham-stan commented 2 years ago

In your code, it seems that the graph is not recomputed in the feature space in each layer. So a static graph is yielded?

yuxumin commented 2 years ago

Yes, we use the original coordinates for graph computing in each layer.