Open basit-7 opened 1 year ago
I understand you mean that whether you want to process your ground truth meshes into a .h5 file as training data or to downsample them into sparse point clouds respectively? If it is the former, I'm sorry that I cannot provide you with a feasible method. Our code is inherited from PU-GCN, the dataset is provided by them. Therefore we donot know how to process them into a .h5 file, and you can ask them for help. For the latter, I think you need to use Poisson disk sampling to get the point clouds with any numbers of points you want.
I wanted to generate training data in .h5 format given 3D meshes. You have provided the scripts here, https://github.com/BingHan0458/PU-GACNet/tree/master/data_preprocessing. I only want to understand the order in which the scripts should be run.
Hi, congratulations on the great work!
I have some ground truth meshes that I need to preprocess to start training. I see two folders in data_preprocessing directory. I am not sure where to start from.
Should the scripts in prep_data be enough? If yes, in what order should I run those scripts.
Where do I need eig_computation_op?