Skip to content

BoxToMask - explicit check for box extending over image shape #6637

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
wyli merged 2 commits into from
Jun 22, 2023
Merged

BoxToMask - explicit check for box extending over image shape #6637

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2ecd048
explicit check for box extending over image shape
kretes Jun 21, 2023
778d4bd
Merge branch 'dev' into box_to_mask_explicit_error
wyli Jun 22, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions monai/apps/detection/transforms/box_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -242,6 +242,9 @@ def convert_box_to_mask(
boxes_mask_np = np.ones((labels.shape[0],) + spatial_size, dtype=np.int16) * np.int16(bg_label)

boxes_np: np.ndarray = convert_data_type(boxes, np.ndarray, dtype=np.int32)[0]
if np.any(boxes_np[:, spatial_dims:] > np.array(spatial_size)):
raise ValueError("Some boxes are larger than the image.")

labels_np, *_ = convert_to_dst_type(src=labels, dst=boxes_np)
for b in range(boxes_np.shape[0]):
# generate a foreground mask
Expand Down
18 changes: 18 additions & 0 deletions tests/test_box_transform.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -131,6 +131,24 @@ def test_value_3d_mask(self):
assert_allclose(data_back["boxes"], data["boxes"], type_test=False, device_test=False, atol=1e-3)
assert_allclose(data_back["labels"], data["labels"], type_test=False, device_test=False, atol=1e-3)

def test_shape_assertion(self):
test_dtype = torch.float32
image = np.zeros((1, 10, 10, 10))
boxes = np.array([[7, 8, 9, 10, 12, 13]])
data = {"image": image, "boxes": boxes, "labels": np.array((1,))}
data = CastToTyped(keys=["image", "boxes"], dtype=test_dtype)(data)
transform_to_mask = BoxToMaskd(
box_keys="boxes",
box_mask_keys="box_mask",
box_ref_image_keys="image",
label_keys="labels",
min_fg_label=0,
ellipse_mask=False,
)
with self.assertRaises(ValueError) as context:
transform_to_mask(data)
self.assertTrue("Some boxes are larger than the image." in str(context.exception))

@parameterized.expand(TESTS_3D)
def test_value_3d(
self,
Expand Down