panel_segmentation.panel_detection.PanelDetection.diceCoeffLoss

PanelDetection.diceCoeffLoss(y_true, y_pred)[source]

This function is a loss function that can be used when training the segmentation model. This loss function can be used in place of binary crossentropy, which is the current loss function in the training stage.

Parameters

  • y_true (numpy array of floats) – The true mask of the image

  • y_pred (numpy array of floats) – The predicted mask of the data

Returns

  • float

  • The loss metric between prediction and ground truth