How to use batchgenerators for regression

[adler-j]

I'm working in inverse problems and quite commonly I have a set of inputs (e.g. noisy volumes) and some outputs (noiseless volume). batchgenerators looks like a perfect fit for data augmentation, but it seems to be designed for image -> segmentation. Is there any way to work with image -> image problems, in that case how do I call the library?

[FabianIsensee]

Hi, you just stack the desired output as an additional color channel in 'data'. Then you make smart use of AppendChannelsTransform. Something like this:

• 'data' is your image information stacked in c with your target output]
• do all spatial transformations, no color transformation!
• use AppendChannelsTransform to move the channel with your target into the key 'seg'. 'seg' will be created automatically (if will give a warning which you can ignore)
• now do color augmentations if applicable (brightness, contrast etc)

voila. That should do it =)

Best, Fabian

[adler-j]

Thank you very much, this makes perfect sense!