search

HuXiaoling / TopoLoss

Code for the NeurIPS 2019 paper: Topology-Preserving Deep Image Segmentation
MIT License
146 stars 22 forks source link

About betti error #21

Closed wlsdzyzl closed 1 year ago

wlsdzyzl commented 1 year ago

Hi, sorry to bother you again. Can you provide the details of computing BETTI error? I saw that in cldice, they computed the betti error on random 64*64 patches. Did you do the experiments under the same setting?

HuXiaoling commented 1 year ago

We did the experiments under the similar setting, while the patch size is slightly different (this should be fine as long as you keep the same patch size for all the baselines). You could also choose the propriate patch size for your own dataset, and the creterion is that the patch should contain useful structures but not complex enough. Also please pay attention to the dimension of betti errors.

wlsdzyzl commented 1 year ago

Ok, thanks.