dongliangcao / unsupervised-learning-of-robust-spectral-shape-matching Goto Github PK

Dear @dongliangcao,

Congratulations for your work and thanks for sharing it with us. I have been following your work regarding the usage of functional maps and I was quite curious to understand whether it could be used for shape assembly.

In brief to understand the problem, you might have two or more pieces of an object which are matching to some part in order to assemble a complete object. So imagine that you have the following pieces of a stone:

These can be assembled together as follows:

So in principle there are specific parts of the objects that should have a bijective matching:

My question now is whether I could use functional maps in order to extract these partial bijective matching parts?

Hello, thank you for your contribution. I am very interested in your research, but I can’t find where the loss of 𝐿couple is defined when reading the code. Can you help me tell me which file it is in.

Hi, thanks for the great work. I am running inference given your provided checkpoints. In Table 3, the result for (train-test) Faust-Faust, Scape-Scape matches with what you have reported. However, the cross dataset F-S, S-F is quite bad. F-S gains 6.66 (2.2 reported), and S-F gains 4.58 (1.6 reported). I leave my config file for F-S below. It would be great if you can take a look to see if there is a problem. Thank you.

# general setting
name: faust_on_scape
backend: dp # DataParallel
type: FMNetModel
num_gpu: auto
manual_seed: 1234
non-isometric: false
partial: false
visualize: true

# path
path:
  resume_state: checkpoints/faust.pth
  resume: false

# datasets
datasets:
  test_dataset:
    name: ScapeVal
    type: PairScapeDataset
    phase: test
    data_root: ../data/SCAPE_r/
    return_evecs: true
    return_faces: true
    num_evecs: 200
    return_corr: true
    return_dist: true

# network setting
networks:
  feature_extractor:
    type: DiffusionNet
    in_channels: 128
    out_channels: 256
    cache_dir: ../data/SCAPE_r/diffusion
    input_type: wks
  permutation:
    type: Similarity
    tau: 0.07
    hard: true

# validation setting
val:
  metrics:
    geo_error:
      type: calculate_geodesic_error
    plot_pck:
      type: plot_pck

Hello @dongliangcao ,

I hope this message finds you well. First of all, I would like to express my appreciation for your work and the valuable contribution you have made to the community.

I have been trying to run your code on a server environment, but unfortunately, I encountered a compatibility issue with the GLIBC version. The server's GLIBC version is lower than the minimum requirement specified in your code. As you may know, upgrading the GLIBC version on the server is not a straightforward task and may have significant implications for other applications running on it.

I wanted to reach out to you to inquire if there are any alternative solutions or workarounds to address this compatibility issue without necessarily upgrading the GLIBC version. I understand that GLIBC is a critical dependency for your code, but I wondered if there might be any configuration options or modifications that could enable successful execution on systems with lower GLIBC versions.

I believe that finding a workaround for this compatibility issue would greatly benefit users who face similar constraints on their server environments. Your insights and suggestions would be highly appreciated.

Thank you for your time and consideration. I look forward to hearing from you and continuing to contribute to the development of your project.

Hi! Your work is amazing! May I ask how to test your work using my own data?
Suppose I have two meshes and want to construct the correspondence between them. How can I make them as the dataset form provided by you?
Thank you!

Dear Author Cao,

I am writing to you regarding your research on Unsupervised Learning of Robust Spectral Shape Matching 2023 ToG. Firstly, I would like to express my appreciation for your contributions to the field.

I have been using the code provided on your GitHub for training on the faust_r dataset. I can replicate the results of FAUST TO FAUST mentioned in your paper. However, upon attempting to conduct generalization testing on the scape_r dataset using the pre-trained final.pth file obtained from the training on the faust_r dataset, I encountered unexpected results.

Specifically, when executing the command python test.py --opt options/test/scape.yaml with the resume_state parameter set to the final.pth file obtained from training on the faust_r dataset, I observed an average error of 0.062. This value significantly deviates from the 0.022 average error mentioned in your paper.

I kindly request guidance on the appropriate steps to take in order to achieve results consistent with those presented in your paper when conducting generalization testing on the scape_r dataset.

Thank you very much for your time and attention to this matter. I look forward to your prompt response.

Best regards,
HJ Xu