CombinedMaskLoss<T>
Models & Types · AiDotNet.ComputerVision.Segmentation.Losses
Combined mask loss using BCE + Dice.
For Beginners
Combining BCE and Dice loss often gives better results than using either alone. BCE provides pixel-level supervision while Dice optimizes the overall mask quality.
Constructors
| Constructor | Summary |
|---|---|
CombinedMaskLoss(double bceWeight, double diceWeight) |
Creates a combined BCE + Dice loss. |
Methods
| Method | Summary |
|---|---|
T Compute(Tensor<T> predicted, Tensor<T> target) |
Computes combined loss. |