Can i ask, now i’m already add mask predict from unet as channel 4 to image RGB, but my macro f1 still around 69%. Now I want to add more feature like age,location,… in fully connected.
thank you so much for your support
Welcome to the forums. There are a lot of unet models on GitHub. Can you link or copy your current model code?
thanks, i can add more feature but my macro still around 85%. Can I ask that is there any other way to increase the macro of the model, now I’m predict benign or malignant using HAM10000.