MaskBCELoss<T>

Models & Types · AiDotNet.ComputerVision.Segmentation.Losses

Binary Cross-Entropy loss for mask prediction.

For Beginners

BCE loss is the standard loss for binary segmentation masks. It computes the cross-entropy between predicted probabilities and binary ground truth.

Constructors

Constructor Summary
MaskBCELoss(double eps) Creates a new mask BCE loss.

Methods

Method Summary
T Compute(Tensor<T> predicted, Tensor<T> target) Computes BCE loss between predicted and target masks.