Regarding changing the network to a single semantic change mask

[Essha123]

Hello,
Thank you for you work and code availabiltiy. I was wondering that is it possible to train your network on a single output label data? I am working on a semantic change detection and i want to compare the results generated by my network with the results generated by your network. However, my network is trained on a change detection dataset with a single change mask for the bitemporal images. Is it possible to change your network to be trained on such dataset? If yes then can you please give me some starting point about how can i change your network?

Once again thank you for your help

[ccstrong007]

你好 可以交流一下吗