search

sw-gong / spiralnet_plus

SpiralNet++: A Fast and Highly Efficient Mesh Convolution Operator (ICCV-W 2019)
https://arxiv.org/abs/1911.05856
MIT License
123 stars 23 forks source link

constant serialization #8

Open jxtxwsbn opened 3 years ago

jxtxwsbn commented 3 years ago

Hi,

Thanks for your code and paper. It looks concise and great. I have a quick question here and hope you could give me some directions.

In the FAUST experiment, the spirals are defined by the first data and keep unchanged during the training. In geometry, this serialization doesn't stand for spirals for other data. Is there any explanation for it?

Thanks!

sw-gong commented 3 years ago

Hi @jxtxwsbn,

The meshes topology in FAUST (as well as datasets used in another two experiments) are the same, which is the underlying requirement to apply spiral on the corresponding task.

jxtxwsbn commented 3 years ago

Thanks so much!

  1. It means faces of different data points in FAUST consist of the same vertex indexes.
  2. The index of the vertice stands for the category it belongs to.
  3. The only difference is the Euclidean coordinate of the vertex.

Do my opinions make sense? Thanks!