about the classification?

[xiaomabufei]

hello, have you considered the classification issue? How do you handle the categories, please?

[tomeramit]

Can you elaborate? I dont understand your question/meaning

[xiaomabufei]

Thank you for your reply! I mean, do you consider the category of mask in your task settings. Do you generate masks for one category at a time or for all categories at once ?

[xiaomabufei]

what the meaning of this phrase? Looking forward to your reply!!

[xiaomabufei]

@tomeramit Can you elaborate on your mission setting? Thank you very much!!

[tomeramit]

The framework is for binary classification while Cityscapes is a semantic segmentation dataset. Therefore, in the paper, we split it into per-object datasets. This means that we created separate datasets for buses, cars, etc., in order to have a binary segmentation dataset. We did this to compare it with the other models in the benchmark that also perform binary segmentation. Does it answer your question?

[tomeramit]

I'm closing the issue, in feel free to reopen in case you want