enhance: support MetaTensor in ExtractDataKeyFromMetaKeyd

- Adds image_only parameter to handle MetaTensor objects
- Prevents redundant metadata extraction when image_only=True
- Improves metadata handling for medical imaging workflows
- Preserves MetaTensor metadata associations
- Addresses issue #7562

Author: skdas20

monai/apps/reconstruction/transforms/dictionary.py

@@ -37,16 +37,19 @@ class ExtractDataKeyFromMetaKeyd(MapTransform):
         keys: keys to be transferred from meta to data
         meta_key: the meta key where all the meta-data is stored
         allow_missing_keys: don't raise exception if key is missing
+        image_only: if True, only extract metadata from MetaTensor images to avoid duplication
 
     Example:
         When the fastMRI dataset is loaded, "kspace" is stored in the data dictionary,
         but the ground-truth image with the key "reconstruction_rss" is stored in the meta data.
         In this case, ExtractDataKeyFromMetaKeyd moves "reconstruction_rss" to data.
+        For MetaTensor objects, setting image_only=True prevents extracting redundant metadata.
     """
 
-    def __init__(self, keys: KeysCollection, meta_key: str, allow_missing_keys: bool = False) -> None:
+    def __init__(self, keys: KeysCollection, meta_key: str, allow_missing_keys: bool = False, image_only: bool = False) -> None:
         MapTransform.__init__(self, keys, allow_missing_keys)
         self.meta_key = meta_key
+        self.image_only = image_only
 
     def __call__(self, data: Mapping[Hashable, NdarrayOrTensor]) -> dict[Hashable, Tensor]:
         """
@@ -60,7 +63,12 @@ def __call__(self, data: Mapping[Hashable, NdarrayOrTensor]) -> dict[Hashable, T
         d = dict(data)
         for key in self.keys:
             if key in d[self.meta_key]:
-                d[key] = d[self.meta_key][key]  # type: ignore
+                extracted_value = d[self.meta_key][key]
+                # When image_only is True, skip if the extracted value is a MetaTensor
+                # to preserve metadata associations
+                if self.image_only and isinstance(extracted_value, MetaTensor):
+                    continue
+                d[key] = extracted_value  # type: ignore
             elif not self.allow_missing_keys:
                 raise KeyError(
                     f"Key `{key}` of transform `{self.__class__.__name__}` was missing in the meta data"