Independent training labels for multiple images

I would like to create a series of simple segmentations using the pixel classifier across a series of images. I have two classes I want to implement. Object and background. Across images the object and background texture and features are likely to differ significantly, however I still want them to have the same label across images.

My current understanding is that the pixel label classifiers across all images contribute cumulatively to final predictions for any individual image. Is there any way to just use the labels within an image to to specify the pixel predictions within the image without having to create separate projects for each image?

Apologies if I missed a post that addresses this issue.

Hi @Oliver_Windram,

sorry, completely missed your post. Answer might come too late for you, but for others it might still help.

In short: no. In order to have individual classifiers, you’d have to create a new project file for each image.