Semantic segmentation NOT overfitting

I am having some problem with a semantic segmentation implementation. I have 2 classes, a background, and an organ. No matter what I do my model only predicts background. Even if I increase the epoch and use training data for a prediction I only get one class as my output. Did anyone see this kind of problem before? Any and all help is appreciated.


Was this issue answered here and can be considered solved?

Yes, It was solved. Thanks for your help