Merge pull request #99 from BrainLesion/more_metrics

More metrics

examples/example_spine_semantic.py

@@ -22,6 +22,7 @@
     expected_input=SemanticPair,
     instance_approximator=ConnectedComponentsInstanceApproximator(),
     instance_matcher=NaiveThresholdMatching(),
+    verbose=True,
 )
 
 with cProfile.Profile() as pr:

panoptica/metrics/metrics.py

@@ -175,6 +175,7 @@ class MetricType(_Enum_Compare):
         _Enum_Compare (_type_): _description_
     """
 
+    NO_PRINT = auto()
     MATCHING = auto()
     GLOBAL = auto()
     INSTANCE = auto()

panoptica/panoptic_evaluator.py

@@ -146,7 +146,6 @@ def panoptic_evaluate(
             instance_approximator is not None
         ), "Got SemanticPair but not InstanceApproximator"
         print("-- Got SemanticPair, will approximate instances")
-        processing_pair = instance_approximator.approximate_instances(processing_pair)
         start = perf_counter()
         processing_pair = instance_approximator.approximate_instances(processing_pair)
         if log_times:
@@ -185,6 +184,7 @@ def panoptic_evaluate(
 
     if isinstance(processing_pair, MatchedInstancePair):
         print("-- Got MatchedInstancePair, will evaluate instances")
+        start = perf_counter()
         processing_pair = evaluate_matched_instance(
             processing_pair,
             eval_metrics=eval_metrics,

panoptica/panoptic_result.py

@@ -13,6 +13,7 @@
     MetricType,
     _compute_centerline_dice_coefficient,
     _compute_dice_coefficient,
+    _average_symmetric_surface_distance,
 )
 from panoptica.utils import EdgeCaseHandler
 
@@ -94,6 +95,20 @@ def __init__(
             fn,
             long_name="False Negatives",
         )
+        self.prec: int
+        self._add_metric(
+            "prec",
+            MetricType.NO_PRINT,
+            prec,
+            long_name="Precision (positive predictive value)",
+        )
+        self.rec: int
+        self._add_metric(
+            "rec",
+            MetricType.NO_PRINT,
+            rec,
+            long_name="Recall (sensitivity)",
+        )
         self.rq: float
         self._add_metric(
             "rq",
@@ -119,6 +134,14 @@ def __init__(
             global_bin_cldsc,
             long_name="Global Binary Centerline Dice",
         )
+        #
+        self.global_bin_assd: int
+        self._add_metric(
+            "global_bin_assd",
+            MetricType.GLOBAL,
+            global_bin_assd,
+            long_name="Global Binary Average Symmetric Surface Distance",
+        )
         # endregion
         #
         # region IOU
@@ -270,6 +293,8 @@ def calculate_all(self, print_errors: bool = False):
     def __str__(self) -> str:
         text = ""
         for metric_type in MetricType:
+            if metric_type == MetricType.NO_PRINT:
+                continue
             text += f"\n+++ {metric_type.name} +++\n"
             for k, v in self._evaluation_metrics.items():
                 if v.metric_type != metric_type:
@@ -360,6 +385,14 @@ def fn(res: PanopticaResult):
     return res.num_ref_instances - res.tp
 
 
+def prec(res: PanopticaResult):
+    return res.tp / (res.tp + res.fp)
+
+
+def rec(res: PanopticaResult):
+    return res.tp / (res.tp + res.fn)
+
+
 def rq(res: PanopticaResult):
     """
     Calculate the Recognition Quality (RQ) based on TP, FP, and FN.
@@ -456,6 +489,16 @@ def global_bin_cldsc(res: PanopticaResult):
     return _compute_centerline_dice_coefficient(ref_binary, pred_binary)
 
 
+def global_bin_assd(res: PanopticaResult):
+    if res.tp == 0:
+        return 0.0
+    pred_binary = res._prediction_arr.copy()
+    ref_binary = res._reference_arr.copy()
+    pred_binary[pred_binary != 0] = 1
+    ref_binary[ref_binary != 0] = 1
+    return _average_symmetric_surface_distance(ref_binary, pred_binary)
+
+
 # endregion