It is defined as a weighted sum of Dice loss and a modified cross entropy. It attempts to leverage the flexibility of Dice loss of class imb

This loss is focuses on less accurately predicted structures using the combination of Dice Loss and Cross Entropy Loss Original pape

More information about this loss, see: https://arxiv.org/pdf/1810.07842.pdf This loss is similar to Tversky Loss, but with a small adjustment

L_{lc-dce} = log(cosh(DiceLoss)

Tversky Loss is the generalization of Dice Loss It in the group of Region-Base Loss