add min overlap ratio as an arg for lesion-wise metrics

MetricsReloaded/metrics/pairwise_measures.py

@@ -247,6 +247,7 @@ def __init__(
         pixdim=None,
         empty=False,
         dict_args={},
+        overlap_ratio=0.1,
     ):
 
         self.measures_dict = {
@@ -300,6 +301,7 @@ def __init__(
         self.connectivity = connectivity_type
         self.pixdim = pixdim
         self.dict_args = dict_args
+        self.overlap_ratio = overlap_ratio
 
     def __fp_map(self):
         """
@@ -1178,7 +1180,9 @@ def measured_hausdorff_distance_perc(self):
     def lesion_wise_tp_fp_fn(self, truth, prediction):
         """
         Computes the true positives, false positives, and false negatives two masks. Masks are considered true positives
-        if at least one voxel overlaps between the truth and the prediction.
+        if there is at least `self.overlap_ratio` overlap between the truth and the prediction.
+        i.e. if self.overlap_ratio = 0.1, then at least 10% of the lesion voxels should overlap between the truth and 
+        the prediction to be considered as true positive.
         Adapted from: https://github.com/npnl/atlas2_grand_challenge/blob/main/isles/scoring.py#L341
 
         Parameters
@@ -1215,7 +1219,7 @@ def lesion_wise_tp_fp_fn(self, truth, prediction):
             num_truth_lesion_voxels = np.sum(lesion)  # Total number of voxels in the GT lesion
             overlapping_voxels = np.sum(lesion * prediction)  # Number of GT voxels that overlap with the prediction
             # Check if at least 10% of the lesion voxels overlap with the prediction
-            if overlapping_voxels / num_truth_lesion_voxels >= 0.1:
+            if overlapping_voxels / num_truth_lesion_voxels >= self.overlap_ratio:
                 tp += 1
             else:
                 fn += 1
@@ -1226,7 +1230,7 @@ def lesion_wise_tp_fp_fn(self, truth, prediction):
             lesion = labeled_prediction == idx_lesion
             # num_pred_lesion_voxels = np.sum(lesion)
             # overlapping_voxels = np.sum(lesion & truth)
-            # if overlapping_voxels / num_pred_lesion_voxels < 0.1:
+            # if overlapping_voxels / num_pred_lesion_voxels < self.overlap_ratio:
             #     fp += 1
             lesion_pred_sum = lesion + truth
             if(np.max(lesion_pred_sum) <= 1):  # No overlap