how to train 2 labels

[adaxidedakaonang]

Hello!

I am new at SegNet and so far I have tested the pretrained models given by deeplab webpage and also trained with my own data. In my first attempt of image segmentation I used an image as input and what I call a segmentation mask as output (binary image).

But now I would like to know if it is possible to have an image as input and as output an image and a depth map. Is it possible to combine these two types of data as "labels"? in other words, can I have two labels as output? I have no idea how to do this and I would really appreciate your help.

Thank you very much!!