[fbsync] Allow SanitizeBoundingBoxes to sanitize more labels (#8319)

Reviewed By: vmoens

Differential Revision: D55062811

fbshipit-source-id: 2216cc6805bff1c69d3dd7e570b80c927b82d19a

Author: NicolasHug

test/test_transforms_v2.py

@@ -5706,7 +5706,17 @@ def _get_boxes_and_valid_mask(self, H=256, W=128, min_size=10):
         return boxes, expected_valid_mask
 
     @pytest.mark.parametrize("min_size", (1, 10))
-    @pytest.mark.parametrize("labels_getter", ("default", lambda inputs: inputs["labels"], None, lambda inputs: None))
+    @pytest.mark.parametrize(
+        "labels_getter",
+        (
+            "default",
+            lambda inputs: inputs["labels"],
+            lambda inputs: (inputs["labels"], inputs["other_labels"]),
+            lambda inputs: [inputs["labels"], inputs["other_labels"]],
+            None,
+            lambda inputs: None,
+        ),
+    )
     @pytest.mark.parametrize("sample_type", (tuple, dict))
     def test_transform(self, min_size, labels_getter, sample_type):
 
@@ -5721,12 +5731,16 @@ def test_transform(self, min_size, labels_getter, sample_type):
 
         labels = torch.arange(boxes.shape[0])
         masks = tv_tensors.Mask(torch.randint(0, 2, size=(boxes.shape[0], H, W)))
+        # other_labels corresponds to properties from COCO like iscrowd, area...
+        # We only sanitize it when labels_getter returns a tuple
+        other_labels = torch.arange(boxes.shape[0])
         whatever = torch.rand(10)
         input_img = torch.randint(0, 256, size=(1, 3, H, W), dtype=torch.uint8)
         sample = {
             "image": input_img,
             "labels": labels,
             "boxes": boxes,
+            "other_labels": other_labels,
             "whatever": whatever,
             "None": None,
             "masks": masks,
@@ -5741,12 +5755,14 @@ def test_transform(self, min_size, labels_getter, sample_type):
         if sample_type is tuple:
             out_image = out[0]
             out_labels = out[1]["labels"]
+            out_other_labels = out[1]["other_labels"]
             out_boxes = out[1]["boxes"]
             out_masks = out[1]["masks"]
             out_whatever = out[1]["whatever"]
         else:
             out_image = out["image"]
             out_labels = out["labels"]
+            out_other_labels = out["other_labels"]
             out_boxes = out["boxes"]
             out_masks = out["masks"]
             out_whatever = out["whatever"]
@@ -5757,14 +5773,20 @@ def test_transform(self, min_size, labels_getter, sample_type):
         assert isinstance(out_boxes, tv_tensors.BoundingBoxes)
         assert isinstance(out_masks, tv_tensors.Mask)
 
-        if labels_getter is None or (callable(labels_getter) and labels_getter({"labels": "blah"}) is None):
+        if labels_getter is None or (callable(labels_getter) and labels_getter(sample) is None):
             assert out_labels is labels
+            assert out_other_labels is other_labels
         else:
             assert isinstance(out_labels, torch.Tensor)
             assert out_boxes.shape[0] == out_labels.shape[0] == out_masks.shape[0]
             # This works because we conveniently set labels to arange(num_boxes)
             assert out_labels.tolist() == valid_indices
 
+            if callable(labels_getter) and isinstance(labels_getter(sample), (tuple, list)):
+                assert_equal(out_other_labels, out_labels)
+            else:
+                assert_equal(out_other_labels, other_labels)
+
     @pytest.mark.parametrize("input_type", (torch.Tensor, tv_tensors.BoundingBoxes))
     def test_functional(self, input_type):
         # Note: the "functional" F.sanitize_bounding_boxes was added after the class, so there is some

torchvision/transforms/v2/_misc.py

@@ -321,6 +321,9 @@ class SanitizeBoundingBoxes(Transform):
     - have any coordinate outside of their corresponding image. You may want to
       call :class:`~torchvision.transforms.v2.ClampBoundingBoxes` first to avoid undesired removals.
 
+    It can also sanitize other tensors like the "iscrowd" or "area" properties from COCO
+    (see ``labels_getter`` parameter).
+
     It is recommended to call it at the end of a pipeline, before passing the
     input to the models. It is critical to call this transform if
     :class:`~torchvision.transforms.v2.RandomIoUCrop` was called.
@@ -330,18 +333,26 @@ class SanitizeBoundingBoxes(Transform):
 
     Args:
         min_size (float, optional) The size below which bounding boxes are removed. Default is 1.
-        labels_getter (callable or str or None, optional): indicates how to identify the labels in the input.
+        labels_getter (callable or str or None, optional): indicates how to identify the labels in the input
+            (or anything else that needs to be sanitized along with the bounding boxes).
             By default, this will try to find a "labels" key in the input (case-insensitive), if
             the input is a dict or it is a tuple whose second element is a dict.
             This heuristic should work well with a lot of datasets, including the built-in torchvision datasets.
-            It can also be a callable that takes the same input
-            as the transform, and returns the labels.
+
+            It can also be a callable that takes the same input as the transform, and returns either:
+
+            - A single tensor (the labels)
+            - A tuple/list of tensors, each of which will be subject to the same sanitization as the bounding boxes.
+              This is useful to sanitize multiple tensors like the labels, and the "iscrowd" or "area" properties
+              from COCO.
+
+            If ``labels_getter`` is None then only bounding boxes are sanitized.
     """
 
     def __init__(
         self,
         min_size: float = 1.0,
-        labels_getter: Union[Callable[[Any], Optional[torch.Tensor]], str, None] = "default",
+        labels_getter: Union[Callable[[Any], Any], str, None] = "default",
     ) -> None:
         super().__init__()
 
@@ -356,18 +367,28 @@ def forward(self, *inputs: Any) -> Any:
         inputs = inputs if len(inputs) > 1 else inputs[0]
 
         labels = self._labels_getter(inputs)
-        if labels is not None and not isinstance(labels, torch.Tensor):
-            raise ValueError(
-                f"The labels in the input to forward() must be a tensor or None, got {type(labels)} instead."
-            )
+        if labels is not None:
+            msg = "The labels in the input to forward() must be a tensor or None, got {type} instead."
+            if isinstance(labels, torch.Tensor):
+                labels = (labels,)
+            elif isinstance(labels, (tuple, list)):
+                for entry in labels:
+                    if not isinstance(entry, torch.Tensor):
+                        # TODO: we don't need to enforce tensors, just that entries are indexable as t[bool_mask]
+                        raise ValueError(msg.format(type=type(entry)))
+            else:
+                raise ValueError(msg.format(type=type(labels)))
 
         flat_inputs, spec = tree_flatten(inputs)
         boxes = get_bounding_boxes(flat_inputs)
 
-        if labels is not None and boxes.shape[0] != labels.shape[0]:
-            raise ValueError(
-                f"Number of boxes (shape={boxes.shape}) and number of labels (shape={labels.shape}) do not match."
-            )
+        if labels is not None:
+            for label in labels:
+                if boxes.shape[0] != label.shape[0]:
+                    raise ValueError(
+                        f"Number of boxes (shape={boxes.shape}) and must match the number of labels."
+                        f"Found labels with shape={label.shape})."
+                    )
 
         valid = F._misc._get_sanitize_bounding_boxes_mask(
             boxes,
@@ -381,7 +402,7 @@ def forward(self, *inputs: Any) -> Any:
         return tree_unflatten(flat_outputs, spec)
 
     def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
-        is_label = inpt is not None and inpt is params["labels"]
+        is_label = params["labels"] is not None and any(inpt is label for label in params["labels"])
         is_bounding_boxes_or_mask = isinstance(inpt, (tv_tensors.BoundingBoxes, tv_tensors.Mask))
 
         if not (is_label or is_bounding_boxes_or_mask):
@@ -391,5 +412,5 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
 
         if is_label:
             return output
-
-        return tv_tensors.wrap(output, like=inpt)
+        else:
+            return tv_tensors.wrap(output, like=inpt)

torchvision/transforms/v2/_utils.py

@@ -4,7 +4,7 @@
 import numbers
 from contextlib import suppress
 
-from typing import Any, Callable, Dict, List, Literal, Optional, Sequence, Tuple, Type, Union
+from typing import Any, Callable, Dict, List, Literal, Sequence, Tuple, Type, Union
 
 import PIL.Image
 import torch
@@ -139,9 +139,7 @@ def _find_labels_default_heuristic(inputs: Any) -> torch.Tensor:
     return inputs[candidate_key]
 
 
-def _parse_labels_getter(
-    labels_getter: Union[str, Callable[[Any], Optional[torch.Tensor]], None]
-) -> Callable[[Any], Optional[torch.Tensor]]:
+def _parse_labels_getter(labels_getter: Union[str, Callable[[Any], Any], None]) -> Callable[[Any], Any]:
     if labels_getter == "default":
         return _find_labels_default_heuristic
     elif callable(labels_getter):