search

mabaorui / NeuralPull

Implementation of ICML'2021:Neural-Pull: Learning Signed Distance Functions from Point Clouds by Learning to Pull Space onto Surfaces
MIT License
179 stars 28 forks source link

Surface reconstruction #5

Closed ThibaultGROUEIX closed 2 years ago

ThibaultGROUEIX commented 2 years ago

Hi,

Thanks a lot for the code release, I greatly enjoyed your paper. I would like to experiment with surface reconstruction on my own pointclouds and ran into a few issues.

As per my understanding from your pape, NeuralPull does not condition the surface reconstruction on anything : 

We do not leverage any condition c in Fig. 1 and overfit the neural network
to the shape by minimizing the loss in Eq. 2

However, the surface reconstruction script seems to be conditioning the reconstruction on a code, and learning surface reconstruction jointly?

Ideally, to test NeuralPull, it would be great if the repo included a script that takes a pointcloud as input, fits neuralPull and returns a mesh.

I am sorry if I misunderstood the code and thanks a lot for your help,

Thibault

mabaorui commented 2 years ago

If you want to input a point cloud and run the script to return a mesh. There are three steps you need to follow:

  1. Prepare data from the input point cloud. python sample_query_point --out_dir ./output_data/ --CUDA 0 --dataset other --input_dir ./input_pointcloud/ Note:The input point cloud should be put into ./input_pointcloud/ folder.
  2. Use the data prepared in step 1 to train the network. python NeuralPull.py --data_dir ./output_data/ --out_dir ./output_network/ --class_idx 02691156 --train --dataset other
  3. Test the trained network in step 2 and output the mesh. python NeuralPull.py --data_dir ./output_data/ --out_dir ./output_network/ --class_idx 02691156 --dataset other

After completing the above three steps, you will get the reconstructed mesh model in the ./output_network/ folder. Thank you for your attention, please contact me if you have any questions.

ThibaultGROUEIX commented 2 years ago

Thanks a lot for getting back to me @mabaorui, I will try these steps.

mabaorui commented 2 years ago

Thanks a lot for getting back to me @mabaorui, I will try these steps.

If you still have any questions, please feel free to contact me.

groueix commented 2 years ago

Thanks for your help, I was able to reproduce your results, Best regards, Thibault