fix non-adaptive thresholding bug (#152)

Author: djdameln

anomalib/models/components/base/anomaly_module.py

@@ -58,10 +58,12 @@ def __init__(self, params: Union[DictConfig, ListConfig]):
         self.model: nn.Module
 
         # metrics
-        auroc = AUROC(num_classes=1, pos_label=1, compute_on_step=False)
-        f1_score = F1(num_classes=1, compute_on_step=False)
-        self.image_metrics = MetricCollection([auroc, f1_score], prefix="image_").cpu()
-        self.pixel_metrics = self.image_metrics.clone(prefix="pixel_").cpu()
+        image_auroc = AUROC(num_classes=1, pos_label=1, compute_on_step=False)
+        image_f1 = F1(num_classes=1, compute_on_step=False, threshold=self.hparams.model.threshold.image_default)
+        pixel_auroc = AUROC(num_classes=1, pos_label=1, compute_on_step=False)
+        pixel_f1 = F1(num_classes=1, compute_on_step=False, threshold=self.hparams.model.threshold.pixel_default)
+        self.image_metrics = MetricCollection([image_auroc, image_f1], prefix="image_").cpu()
+        self.pixel_metrics = MetricCollection([pixel_auroc, pixel_f1], prefix="pixel_").cpu()
 
     def forward(self, batch):  # pylint: disable=arguments-differ
         """Forward-pass input tensor to the module.

tests/pre_merge/utils/metrics/test_adaptive_threshold.py

@@ -13,10 +13,16 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions
 # and limitations under the License.
+import random
 
 import pytest
 import torch
+from pytorch_lightning import Trainer
 
+from anomalib.config import get_configurable_parameters
+from anomalib.data import get_datamodule
+from anomalib.models import get_model
+from anomalib.utils.callbacks import get_callbacks
 from anomalib.utils.metrics import AdaptiveThreshold
 
 
@@ -35,3 +41,29 @@ def test_adaptive_threshold(labels, preds, target_threshold):
     threshold_value = adaptive_threshold.compute()
 
     assert threshold_value == target_threshold
+
+
+def test_non_adaptive_threshold():
+    """
+    Test if the non-adaptive threshold gets used in the F1 score computation when
+    adaptive thresholding is disabled and no normalization is used.
+    """
+    config = get_configurable_parameters(model_config_path="anomalib/models/padim/config.yaml")
+
+    config.model.normalization_method = "none"
+    config.model.threshold.adaptive = False
+    config.trainer.fast_dev_run = True
+
+    image_threshold = random.random()
+    pixel_threshold = random.random()
+    config.model.threshold.image_default = image_threshold
+    config.model.threshold.pixel_default = pixel_threshold
+
+    model = get_model(config)
+    datamodule = get_datamodule(config)
+    callbacks = get_callbacks(config)
+
+    trainer = Trainer(**config.trainer, callbacks=callbacks)
+    trainer.fit(model=model, datamodule=datamodule)
+    assert trainer.model.image_metrics.F1.threshold == image_threshold
+    assert trainer.model.pixel_metrics.F1.threshold == pixel_threshold