diff --git a/src/depiction/image/multi_channel_image.py b/src/depiction/image/multi_channel_image.py
index 84bd052..e401335 100644
--- a/src/depiction/image/multi_channel_image.py
+++ b/src/depiction/image/multi_channel_image.py
@@ -9,8 +9,8 @@
 from xarray import DataArray
 
 from depiction.image.image_channel_stats import ImageChannelStats
-from depiction.persistence.image.hdf5_image_format import Hdf5ImageFormat
 from depiction.image.sparse_representation import SparseRepresentation
+from depiction.persistence.image.hdf5_image_format import Hdf5ImageFormat
 
 if TYPE_CHECKING:
     from collections.abc import Sequence
@@ -157,8 +157,14 @@ def bg_mask_flat(self) -> DataArray:
     def dimensions(self) -> tuple[int, int]:
         """Returns width and height of the image."""
         # TODO reconsider this method (adding it now for compatibility)
+        warnings.warn("dimensions is deprecated, use sizes instead", DeprecationWarning)
         return self._data.sizes["x"], self._data.sizes["y"]
 
+    @property
+    def sizes(self) -> dict[Literal["y", "x", "c"], int]:
+        """Size of the image along each dimension"""
+        return {k: self._data.sizes[k] for k in ["y", "x", "c"]}
+
     @property
     def channel_names(self) -> list[str]:
         """Returns the names of the channels."""
diff --git a/tests/unit/image/test_multi_channel_image.py b/tests/unit/image/test_multi_channel_image.py
index 31b2d37..a9aefb8 100644
--- a/tests/unit/image/test_multi_channel_image.py
+++ b/tests/unit/image/test_multi_channel_image.py
@@ -124,6 +124,10 @@ def test_dimensions(mock_image: MultiChannelImage) -> None:
     assert mock_image.dimensions == (2, 3)
 
 
+def test_sizes(mock_image: MultiChannelImage) -> None:
+    assert mock_image.sizes == {"y": 3, "x": 2, "c": 2}
+
+
 def test_channel_names_when_set(mock_image: MultiChannelImage) -> None:
     # TODO there should be some functionality to make it work for on-the-fly generated channel names
     assert mock_image.channel_names == ["Channel A", "Channel B"]