Align_corners switch in Volumes

Summary:
Porting this commit by davnov134 .
93a3a62800 (diff-a8e107ebe039de52ca112ac6ddfba6ebccd53b4f53030b986e13f019fe57a378)

Capability to interpret world/local coordinates with various align_corners semantics.

Reviewed By: bottler

Differential Revision: D51855420

fbshipit-source-id: 834cd220c25d7f0143d8a55ba880da5977099dd6
This commit is contained in:
Roman Shapovalov 2023-12-07 03:07:41 -08:00 committed by Facebook GitHub Bot
parent fbc6725f03
commit 94da8841af
5 changed files with 91 additions and 9 deletions

View File

@ -98,6 +98,13 @@ def save_model(model, stats, fl, optimizer=None, cfg=None):
return flstats, flmodel, flopt return flstats, flmodel, flopt
def save_stats(stats, fl, cfg=None):
flstats = get_stats_path(fl)
logger.info("saving model stats to %s" % flstats)
stats.save(flstats)
return flstats
def load_model(fl, map_location: Optional[dict]): def load_model(fl, map_location: Optional[dict]):
flstats = get_stats_path(fl) flstats = get_stats_path(fl)
flmodel = get_model_path(fl) flmodel = get_model_path(fl)

View File

@ -291,6 +291,7 @@ def add_pointclouds_to_volumes(
mask=mask, mask=mask,
mode=mode, mode=mode,
rescale_features=rescale_features, rescale_features=rescale_features,
align_corners=initial_volumes.get_align_corners(),
_python=_python, _python=_python,
) )
@ -310,6 +311,7 @@ def add_points_features_to_volume_densities_features(
grid_sizes: Optional[torch.LongTensor] = None, grid_sizes: Optional[torch.LongTensor] = None,
rescale_features: bool = True, rescale_features: bool = True,
_python: bool = False, _python: bool = False,
align_corners: bool = True,
) -> Tuple[torch.Tensor, torch.Tensor]: ) -> Tuple[torch.Tensor, torch.Tensor]:
""" """
Convert a batch of point clouds represented with tensors of per-point Convert a batch of point clouds represented with tensors of per-point
@ -356,6 +358,7 @@ def add_points_features_to_volume_densities_features(
output densities are just summed without rescaling, so output densities are just summed without rescaling, so
you may need to rescale them afterwards. you may need to rescale them afterwards.
_python: Set to True to use a pure Python implementation. _python: Set to True to use a pure Python implementation.
align_corners: as for grid_sample.
Returns: Returns:
volume_features: Output volume of shape `(minibatch, feature_dim, D, H, W)` volume_features: Output volume of shape `(minibatch, feature_dim, D, H, W)`
volume_densities: Occupancy volume of shape `(minibatch, 1, D, H, W)` volume_densities: Occupancy volume of shape `(minibatch, 1, D, H, W)`
@ -409,7 +412,7 @@ def add_points_features_to_volume_densities_features(
grid_sizes, grid_sizes,
1.0, # point_weight 1.0, # point_weight
mask, mask,
True, # align_corners align_corners, # align_corners
splat, splat,
) )

View File

@ -382,9 +382,9 @@ class VolumeSampler(torch.nn.Module):
rays_densities = torch.nn.functional.grid_sample( rays_densities = torch.nn.functional.grid_sample(
volumes_densities, volumes_densities,
rays_points_local_flat, rays_points_local_flat,
align_corners=True,
mode=self._sample_mode, mode=self._sample_mode,
padding_mode=self._padding_mode, padding_mode=self._padding_mode,
align_corners=self._volumes.get_align_corners(),
) )
# permute the dimensions & reshape densities after sampling # permute the dimensions & reshape densities after sampling
@ -400,9 +400,9 @@ class VolumeSampler(torch.nn.Module):
rays_features = torch.nn.functional.grid_sample( rays_features = torch.nn.functional.grid_sample(
volumes_features, volumes_features,
rays_points_local_flat, rays_points_local_flat,
align_corners=True,
mode=self._sample_mode, mode=self._sample_mode,
padding_mode=self._padding_mode, padding_mode=self._padding_mode,
align_corners=self._volumes.get_align_corners(),
) )
# permute the dimensions & reshape features after sampling # permute the dimensions & reshape features after sampling

View File

@ -85,7 +85,7 @@ class Volumes:
are linearly interpolated over the spatial dimensions of the volume. are linearly interpolated over the spatial dimensions of the volume.
- Note that the convention is the same as for the 5D version of the - Note that the convention is the same as for the 5D version of the
`torch.nn.functional.grid_sample` function called with `torch.nn.functional.grid_sample` function called with
`align_corners==True`. the same value of `align_corners` argument.
- Note that the local coordinate convention of `Volumes` - Note that the local coordinate convention of `Volumes`
(+X = left to right, +Y = top to bottom, +Z = away from the user) (+X = left to right, +Y = top to bottom, +Z = away from the user)
is *different* from the world coordinate convention of the is *different* from the world coordinate convention of the
@ -143,7 +143,7 @@ class Volumes:
torch.nn.functional.grid_sample( torch.nn.functional.grid_sample(
v.densities(), v.densities(),
v.get_coord_grid(world_coordinates=False), v.get_coord_grid(world_coordinates=False),
align_corners=True, align_corners=align_corners,
) == v.densities(), ) == v.densities(),
i.e. sampling the volume at trivial local coordinates i.e. sampling the volume at trivial local coordinates
@ -157,6 +157,7 @@ class Volumes:
features: Optional[_TensorBatch] = None, features: Optional[_TensorBatch] = None,
voxel_size: _VoxelSize = 1.0, voxel_size: _VoxelSize = 1.0,
volume_translation: _Translation = (0.0, 0.0, 0.0), volume_translation: _Translation = (0.0, 0.0, 0.0),
align_corners: bool = True,
) -> None: ) -> None:
""" """
Args: Args:
@ -186,6 +187,10 @@ class Volumes:
b) a Tensor of shape (3,) b) a Tensor of shape (3,)
c) a Tensor of shape (minibatch, 3) c) a Tensor of shape (minibatch, 3)
d) a Tensor of shape (1,) (square voxels) d) a Tensor of shape (1,) (square voxels)
**align_corners**: If set (default), the coordinates of the corner voxels are
exactly 1 or +1 in the local coordinate system. Otherwise, the coordinates
correspond to the centers of the corner voxels. Cf. the namesake argument to
`torch.nn.functional.grid_sample`.
""" """
# handle densities # handle densities
@ -206,6 +211,7 @@ class Volumes:
voxel_size=voxel_size, voxel_size=voxel_size,
volume_translation=volume_translation, volume_translation=volume_translation,
device=self.device, device=self.device,
align_corners=align_corners,
) )
# handle features # handle features
@ -336,6 +342,13 @@ class Volumes:
return None return None
return self._features_densities_list(features_) return self._features_densities_list(features_)
def get_align_corners(self) -> bool:
"""
Return whether the corners of the voxels should be aligned with the
image pixels.
"""
return self.locator._align_corners
def _features_densities_list(self, x: torch.Tensor) -> List[torch.Tensor]: def _features_densities_list(self, x: torch.Tensor) -> List[torch.Tensor]:
""" """
Retrieve the list representation of features/densities. Retrieve the list representation of features/densities.
@ -576,7 +589,7 @@ class VolumeLocator:
are linearly interpolated over the spatial dimensions of the volume. are linearly interpolated over the spatial dimensions of the volume.
- Note that the convention is the same as for the 5D version of the - Note that the convention is the same as for the 5D version of the
`torch.nn.functional.grid_sample` function called with `torch.nn.functional.grid_sample` function called with
`align_corners==True`. the same value of `align_corners` argument.
- Note that the local coordinate convention of `VolumeLocator` - Note that the local coordinate convention of `VolumeLocator`
(+X = left to right, +Y = top to bottom, +Z = away from the user) (+X = left to right, +Y = top to bottom, +Z = away from the user)
is *different* from the world coordinate convention of the is *different* from the world coordinate convention of the
@ -634,7 +647,7 @@ class VolumeLocator:
torch.nn.functional.grid_sample( torch.nn.functional.grid_sample(
v.densities(), v.densities(),
v.get_coord_grid(world_coordinates=False), v.get_coord_grid(world_coordinates=False),
align_corners=True, align_corners=align_corners,
) == v.densities(), ) == v.densities(),
i.e. sampling the volume at trivial local coordinates i.e. sampling the volume at trivial local coordinates
@ -651,6 +664,7 @@ class VolumeLocator:
device: torch.device, device: torch.device,
voxel_size: _VoxelSize = 1.0, voxel_size: _VoxelSize = 1.0,
volume_translation: _Translation = (0.0, 0.0, 0.0), volume_translation: _Translation = (0.0, 0.0, 0.0),
align_corners: bool = True,
): ):
""" """
**batch_size** : Batch size of the underlying grids **batch_size** : Batch size of the underlying grids
@ -674,15 +688,21 @@ class VolumeLocator:
b) a Tensor of shape (3,) b) a Tensor of shape (3,)
c) a Tensor of shape (minibatch, 3) c) a Tensor of shape (minibatch, 3)
d) a Tensor of shape (1,) (square voxels) d) a Tensor of shape (1,) (square voxels)
**align_corners**: If set (default), the coordinates of the corner voxels are
exactly 1 or +1 in the local coordinate system. Otherwise, the coordinates
correspond to the centers of the corner voxels. Cf. the namesake argument to
`torch.nn.functional.grid_sample`.
""" """
self.device = device self.device = device
self._batch_size = batch_size self._batch_size = batch_size
self._grid_sizes = self._convert_grid_sizes2tensor(grid_sizes) self._grid_sizes = self._convert_grid_sizes2tensor(grid_sizes)
self._resolution = tuple(torch.max(self._grid_sizes.cpu(), dim=0).values) self._resolution = tuple(torch.max(self._grid_sizes.cpu(), dim=0).values)
self._align_corners = align_corners
# set the local_to_world transform # set the local_to_world transform
self._set_local_to_world_transform( self._set_local_to_world_transform(
voxel_size=voxel_size, volume_translation=volume_translation voxel_size=voxel_size,
volume_translation=volume_translation,
) )
def _convert_grid_sizes2tensor( def _convert_grid_sizes2tensor(
@ -806,8 +826,17 @@ class VolumeLocator:
grid_sizes = self.get_grid_sizes() grid_sizes = self.get_grid_sizes()
# generate coordinate axes # generate coordinate axes
def corner_coord_adjustment(r):
return 0.0 if self._align_corners else 1.0 / r
vol_axes = [ vol_axes = [
torch.linspace(-1.0, 1.0, r, dtype=torch.float32, device=self.device) torch.linspace(
-1.0 + corner_coord_adjustment(r),
1.0 - corner_coord_adjustment(r),
r,
dtype=torch.float32,
device=self.device,
)
for r in (de, he, wi) for r in (de, he, wi)
] ]

View File

@ -312,6 +312,49 @@ class TestVolumes(TestCaseMixin, unittest.TestCase):
).permute(0, 2, 3, 4, 1) ).permute(0, 2, 3, 4, 1)
self.assertClose(grid_world_resampled, grid_world, atol=1e-7) self.assertClose(grid_world_resampled, grid_world, atol=1e-7)
for align_corners in [True, False]:
v_trivial = Volumes(densities=densities, align_corners=align_corners)
# check the case with x_world=(0,0,0)
pts_world = torch.zeros(
num_volumes, 1, 3, device=device, dtype=torch.float32
)
pts_local = v_trivial.world_to_local_coords(pts_world)
pts_local_expected = torch.zeros_like(pts_local)
self.assertClose(pts_local, pts_local_expected)
# check the case with x_world=(-2, 3, -2)
pts_world_tuple = [-2, 3, -2]
pts_world = torch.tensor(
pts_world_tuple, device=device, dtype=torch.float32
)[None, None].repeat(num_volumes, 1, 1)
pts_local = v_trivial.world_to_local_coords(pts_world)
pts_local_expected = torch.tensor(
[-1, 1, -1], device=device, dtype=torch.float32
)[None, None].repeat(num_volumes, 1, 1)
self.assertClose(pts_local, pts_local_expected)
# # check that the central voxel has coords x_world=(0, 0, 0) and x_local(0, 0, 0)
grid_world = v_trivial.get_coord_grid(world_coordinates=True)
grid_local = v_trivial.get_coord_grid(world_coordinates=False)
for grid in (grid_world, grid_local):
x0 = grid[0, :, :, 2, 0]
y0 = grid[0, :, 3, :, 1]
z0 = grid[0, 2, :, :, 2]
for coord_line in (x0, y0, z0):
self.assertClose(
coord_line, torch.zeros_like(coord_line), atol=1e-7
)
# resample grid_world using grid_sampler with local coords
# -> make sure the resampled version is the same as original
grid_world_resampled = torch.nn.functional.grid_sample(
grid_world.permute(0, 4, 1, 2, 3),
grid_local,
align_corners=align_corners,
).permute(0, 2, 3, 4, 1)
self.assertClose(grid_world_resampled, grid_world, atol=1e-7)
def test_coord_grid_convention_heterogeneous( def test_coord_grid_convention_heterogeneous(
self, num_channels=4, dtype=torch.float32 self, num_channels=4, dtype=torch.float32
): ):