search

XuyangBai / PPF-FoldNet

PyTorch reimplementation for "PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptors" https://arxiv.org/abs/1808.10322
99 stars 19 forks source link

Why the number is 1 in your code? #9

Closed roxanasoto closed 2 years ago

roxanasoto commented 3 years ago

Hi, thanks for code! I have a question: why the number of patch is 1 in your code? In the original paper the number of patch es 2048 or 5k, but in training.py you defined in the líne 21 like 1? Thanks!

XuyangBai commented 2 years ago

I might use a misleading variable name during the implementation. To be specific, I use 1 point cloud pairs, select 32 corresponding point cloud patches, each patch contains 1024 points (from the paper the authors use 2048). I forget why I use 1024 instead (maybe due to memory limitation, or the performance is similar)

https://github.com/XuyangBai/PPF-FoldNet/blob/63af940632ad2350141e332314aab99f41766378/train.py#L21-L23