AC: allow resize if impossible extend mask (#2918)

Author: eaidova

tools/accuracy_checker/openvino/tools/accuracy_checker/postprocessor/extend_segmentation_mask.py

@@ -20,8 +20,9 @@
 from .postprocessor import Postprocessor
 from .resize_segmentation_mask import ResizeSegmentationMask
 from ..representation import SegmentationAnnotation, SegmentationPrediction
-from ..config import NumberField, ConfigError, StringField
+from ..config import NumberField, StringField
 from ..preprocessor.geometric_transformations import padding_func
+from ..logging import warning
 
 
 class ExtendSegmentationMask(Postprocessor):
@@ -58,7 +59,10 @@ def process_image(self, annotation, prediction):
             dst_height, dst_width = prediction_.mask.shape[-2:]
             height, width = annotation_mask.shape[-2:]
             if dst_width < width or dst_height < height:
-                raise ConfigError('size for extending should be not less current mask size')
+                warning('size for extending should be not less current mask size. resize operation will be applied')
+                annotation_.mask = self._resize(annotation_.mask, dst_height, dst_width, False)
+                continue
+
 
             pad = self.pad_func(dst_width, dst_height, width, height)
             extended_mask = cv2.copyMakeBorder(
@@ -75,18 +79,6 @@ def process_image_with_metadata(self, annotation, prediction, image_metadata=Non
 
     @staticmethod
     def _deprocess_prediction(prediction, meta):
-        def _resize(entry, height, width):
-            if len(entry.shape) == 2:
-                entry = ResizeSegmentationMask.segm_resize(entry, width, height)
-                return entry
-
-            entry_mask = []
-            for class_mask in entry:
-                resized_mask = ResizeSegmentationMask.segm_resize(class_mask, width, height)
-                entry_mask.append(resized_mask)
-            entry = np.array(entry_mask)
-
-            return entry
         geom_ops = meta.get('geometric_operations', [])
         pad = geom_ops[-1].parameters['pad'] if geom_ops and geom_ops[-1].type == 'padding' else [0, 0, 0, 0]
         image_h, image_w = meta['image_size'][:2]
@@ -96,7 +88,21 @@ def _resize(entry, height, width):
             pred_mask = prediction_.mask[pad[0]:pred_w-pad[2], pad[1]:pred_h-pad[3]]
             pred_h, pred_w = pred_mask.shape[-2:]
             if (pred_h, pred_w) != (image_h, image_w):
-                pred_mask = _resize(pred_mask, image_h, image_w)
+                pred_mask = ExtendSegmentationMask._resize(pred_mask, image_h, image_w)
             prediction_.mask = pred_mask
 
         return prediction
+
+    @staticmethod
+    def _resize(entry, height, width, per_class=True):
+        if len(entry.shape) == 2 or not per_class:
+            entry = ResizeSegmentationMask.segm_resize(entry, width, height)
+            return entry
+
+        entry_mask = []
+        for class_mask in entry:
+            resized_mask = ResizeSegmentationMask.segm_resize(class_mask, width, height)
+            entry_mask.append(resized_mask)
+        entry = np.array(entry_mask)
+
+        return entry

tools/accuracy_checker/tests/test_postprocessor.py

@@ -996,13 +996,6 @@ def test_extend_segmentation_mask_asymmetrical(self):
         assert np.array_equal(prediction[0].mask, expected_prediction_mask)
         assert np.array_equal(annotation[0].mask, expected_annotation_mask)
 
-    def test_extend_segmentation_mask_raise_config_error_if_prediction_less_annotation(self):
-        config = [{'type': 'extend_segmentation_mask'}]
-        annotation = make_segmentation_representation(np.zeros((5, 5)), ground_truth=True)
-        prediction = make_segmentation_representation(np.zeros((4, 4)), ground_truth=False)
-        with pytest.raises(ConfigError):
-            postprocess_data(PostprocessingExecutor(config), annotation, prediction)
-
     def test_extend_segmentation_mask_with_filling_label(self):
         config = [{'type': 'extend_segmentation_mask', 'filling_label': 1}]
         annotation = make_segmentation_representation(np.zeros((5, 5)), ground_truth=True)