5381 and deprecate compute_meandice compute_meaniou (#5382)

Signed-off-by: Wenqi Li <[email protected]>

Fixes #5381

- non-breaking changes to rename `compute_meandice` to `compute_dice`
- non-breaking changes to rename `compute_meaniou` to `compute_iou`

<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Non-breaking change (fix or new feature that would not break
existing functionality).
- [ ] Breaking change (fix or new feature that would cause existing
functionality to change).
- [x] New tests added to cover the changes.
- [ ] Integration tests passed locally by running `./runtests.sh -f -u
--net --coverage`.
- [x] Quick tests passed locally by running `./runtests.sh --quick
--unittests --disttests`.
- [ ] In-line docstrings updated.
- [ ] Documentation updated, tested `make html` command in the `docs/`
folder.

Signed-off-by: Wenqi Li <[email protected]>

CHANGELOG.md

@@ -36,6 +36,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 * Automatically infer device from the first item in random elastic deformation dict
 * Add channel dim in `ComputeHoVerMaps` and `ComputeHoVerMapsd`
 * Remove batch dim in `SobelGradients` and `SobelGradientsd`
+### Deprecated
+* Deprecating `compute_meandice`, `compute_meaniou` in `monai.metrics`, in favor of
+`compute_dice` and `compute_iou` respectively
 
 ## [1.0.0] - 2022-09-16
 ### Added

monai/metrics/__init__.py

@@ -15,8 +15,8 @@
 from .froc import compute_fp_tp_probs, compute_froc_curve_data, compute_froc_score
 from .generalized_dice import GeneralizedDiceScore, compute_generalized_dice
 from .hausdorff_distance import HausdorffDistanceMetric, compute_hausdorff_distance, compute_percent_hausdorff_distance
-from .meandice import DiceMetric, compute_meandice
-from .meaniou import MeanIoU, compute_meaniou
+from .meandice import DiceMetric, compute_dice, compute_meandice
+from .meaniou import MeanIoU, compute_iou, compute_meaniou
 from .metric import Cumulative, CumulativeIterationMetric, IterationMetric, Metric
 from .regression import MAEMetric, MSEMetric, PSNRMetric, RMSEMetric, SSIMMetric
 from .rocauc import ROCAUCMetric, compute_roc_auc

monai/metrics/active_learning_metrics.py

@@ -22,7 +22,7 @@
 
 class VarianceMetric(Metric):
     """
-    Compute the Variance of a given T-repeats N-dimensional array/tensor. The primary usage is as a uncertainty based
+    Compute the Variance of a given T-repeats N-dimensional array/tensor. The primary usage is as an uncertainty based
     metric for Active Learning.
 
     It can return the spatial variance/uncertainty map based on user choice or a single scalar value via mean/sum of the

monai/metrics/meandice.py

@@ -14,7 +14,7 @@
 import torch
 
 from monai.metrics.utils import do_metric_reduction, ignore_background, is_binary_tensor
-from monai.utils import MetricReduction
+from monai.utils import MetricReduction, deprecated
 
 from .metric import CumulativeIterationMetric
 
@@ -80,7 +80,7 @@ def _compute_tensor(self, y_pred: torch.Tensor, y: torch.Tensor):  # type: ignor
         if dims < 3:
             raise ValueError(f"y_pred should have at least 3 dimensions (batch, channel, spatial), got {dims}.")
         # compute dice (BxC) for each channel for each batch
-        return compute_meandice(
+        return compute_dice(
             y_pred=y_pred, y=y, include_background=self.include_background, ignore_empty=self.ignore_empty
         )
 
@@ -103,10 +103,10 @@ def aggregate(self, reduction: Union[MetricReduction, str, None] = None):
         return (f, not_nans) if self.get_not_nans else f
 
 
-def compute_meandice(
+def compute_dice(
     y_pred: torch.Tensor, y: torch.Tensor, include_background: bool = True, ignore_empty: bool = True
 ) -> torch.Tensor:
-    """Computes Dice score metric from full size Tensor and collects average.
+    """Computes Dice score metric for a batch of predictions.
 
     Args:
         y_pred: input data to compute, typical segmentation model output.
@@ -146,6 +146,11 @@ def compute_meandice(
     y_pred_o = torch.sum(y_pred, dim=reduce_axis)
     denominator = y_o + y_pred_o
 
-    if ignore_empty is True:
+    if ignore_empty:
         return torch.where(y_o > 0, (2.0 * intersection) / denominator, torch.tensor(float("nan"), device=y_o.device))
     return torch.where(denominator > 0, (2.0 * intersection) / denominator, torch.tensor(1.0, device=y_o.device))
+
+
+@deprecated(since="1.0.0", msg_suffix="use `compute_dice` instead.")
+def compute_meandice(*args, **kwargs):
+    return compute_dice(*args, **kwargs)

monai/metrics/meaniou.py

@@ -14,14 +14,15 @@
 import torch
 
 from monai.metrics.utils import do_metric_reduction, ignore_background, is_binary_tensor
-from monai.utils import MetricReduction
+from monai.utils import MetricReduction, deprecated
 
 from .metric import CumulativeIterationMetric
 
 
 class MeanIoU(CumulativeIterationMetric):
     """
-    Compute average IoU score between two tensors. It can support both multi-classes and multi-labels tasks.
+    Compute average Intersection over Union (IoU) score between two tensors.
+    It supports both multi-classes and multi-labels tasks.
     Input `y_pred` is compared with ground truth `y`.
     `y_pred` is expected to have binarized predictions and `y` should be in one-hot format. You can use suitable transforms
     in ``monai.transforms.post`` first to achieve binarized values.
@@ -80,7 +81,7 @@ def _compute_tensor(self, y_pred: torch.Tensor, y: torch.Tensor):  # type: ignor
         if dims < 3:
             raise ValueError(f"y_pred should have at least 3 dimensions (batch, channel, spatial), got {dims}.")
         # compute IoU (BxC) for each channel for each batch
-        return compute_meaniou(
+        return compute_iou(
             y_pred=y_pred, y=y, include_background=self.include_background, ignore_empty=self.ignore_empty
         )
 
@@ -103,10 +104,10 @@ def aggregate(self, reduction: Union[MetricReduction, str, None] = None):
         return (f, not_nans) if self.get_not_nans else f
 
 
-def compute_meaniou(
+def compute_iou(
     y_pred: torch.Tensor, y: torch.Tensor, include_background: bool = True, ignore_empty: bool = True
 ) -> torch.Tensor:
-    """Computes IoU score metric from full size Tensor and collects average.
+    """Computes Intersection over Union (IoU) score metric from a batch of predictions.
 
     Args:
         y_pred: input data to compute, typical segmentation model output.
@@ -146,6 +147,11 @@ def compute_meaniou(
     y_pred_o = torch.sum(y_pred, dim=reduce_axis)
     union = y_o + y_pred_o - intersection
 
-    if ignore_empty is True:
+    if ignore_empty:
         return torch.where(y_o > 0, (intersection) / union, torch.tensor(float("nan"), device=y_o.device))
     return torch.where(union > 0, (intersection) / union, torch.tensor(1.0, device=y_o.device))
+
+
+@deprecated(since="1.0.0", msg_suffix="use `compute_iou` instead.")
+def compute_meaniou(*args, **kwargs):
+    return compute_iou(*args, **kwargs)

tests/test_compute_meandice.py

@@ -15,7 +15,7 @@
 import torch
 from parameterized import parameterized
 
-from monai.metrics import DiceMetric, compute_meandice
+from monai.metrics import DiceMetric, compute_dice, compute_meandice
 
 _device = "cuda:0" if torch.cuda.is_available() else "cpu"
 # keep background
@@ -187,7 +187,7 @@
 class TestComputeMeanDice(unittest.TestCase):
     @parameterized.expand([TEST_CASE_1, TEST_CASE_2, TEST_CASE_9, TEST_CASE_11, TEST_CASE_12])
     def test_value(self, input_data, expected_value):
-        result = compute_meandice(**input_data)
+        result = compute_dice(**input_data)
         np.testing.assert_allclose(result.cpu().numpy(), expected_value, atol=1e-4)
 
     @parameterized.expand([TEST_CASE_3])

tests/test_compute_meaniou.py

@@ -15,7 +15,7 @@
 import torch
 from parameterized import parameterized
 
-from monai.metrics import MeanIoU, compute_meaniou
+from monai.metrics import MeanIoU, compute_iou, compute_meaniou
 
 _device = "cuda:0" if torch.cuda.is_available() else "cpu"
 # keep background
@@ -192,7 +192,7 @@ def test_value(self, input_data, expected_value):
 
     @parameterized.expand([TEST_CASE_3])
     def test_nans(self, input_data, expected_value):
-        result = compute_meaniou(**input_data)
+        result = compute_iou(**input_data)
         self.assertTrue(np.allclose(np.isnan(result.cpu().numpy()), expected_value))
 
     # MeanIoU class tests