search

antao97 / dgcnn.pytorch

A PyTorch implementation of Dynamic Graph CNN for Learning on Point Clouds (DGCNN)
MIT License
715 stars 151 forks source link

3d coordinates + extra feature #81

Open AlexandreDiPiazza opened 1 year ago

AlexandreDiPiazza commented 1 year ago

Hello,

Is it possible to use your model for classification, but instead of having 3d inputs (x,y,z), I have 4d inputs (x,y,z,label). Thank you and have a nice day

antao97 commented 1 year ago

Sorry, I don't understand how your classification works.

In this repo, the model can do the classification task. The input is a point cloud of $N$ points in XYZ coordinate (size $N \times 3$). The output is a label vector (size $K$) where each element shows the confidence of a specific class in total $K$ classes.