Recall and precission

Recall is calculated as the ratio of the number of foreground pixels correctly identified to the number of foreground pixels in the ground truth and precision is computed as the ratio of the number of foreground pixels correctly identi¯ed to the number of foreground pixels detected.

J. Davis and M. Goadrich. The relationship between precision-recall and ROC curves. In Proc. 23rd Intl. Conf. on Machine Learning,

2006.