John1231983
John1231983
/ jaccard_coef_loss.py
Created
March 9, 2018 14:34
— forked from wassname/jaccard_coef_loss.py
jaccard_coef_loss for keras. This loss is usefull when you have unbalanced classes within a sample such as segmenting each pixel of an image. For example you are trying to predict if each pixel is cat, dog, or background. You may have 80% background, 10% dog, and 10% cat. Should a model that predicts 100% background be 80% right, or 30%? Categor…
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| from keras import backend as K | |
| def jaccard_distance_loss(y_true, y_pred, smooth=100): | |
| """ | |
| Jaccard = (|X & Y|)/ (|X|+ |Y| - |X & Y|) | |
| = sum(|A*B|)/(sum(|A|)+sum(|B|)-sum(|A*B|)) | |
| The jaccard distance loss is usefull for unbalanced datasets. This has been | |
| shifted so it converges on 0 and is smoothed to avoid exploding or disapearing | |
| gradient. |