autometric is to help draw ROC,PR curves for image semantic segmentation problem.When you have groundtruth image and
predict probabilistic image,this package can draw ROC,PR curves and save them to your savepath.It can also calculate FPR
(False Positive Rate) TPR(True Positive Rate) AUC(Area under Curve) MAP(Mean Average Precision) IoU(Intersection over Union)
at the same time.
Now you can use pip to install autometric,try pip install autometric
requirements:
- python 3.0+
- opencv_python
- matplotlib
- numpy
autometric include the following function/class:
CollectData: aclassto calculate FPR,TPR,AUC,MAP,IOUDrawCurve: aclassto draw and save ROC PR curvesfileList: afunctionto collect file to a listdrawCurve: afunctionto draw curves and calculate parameters.Read example.py to get an idea of how to use!
if you right use the package,you can see the roc curve like below:
