Skip to content

proc_ops ¤

Classes:

Name Description
AddKnownDatasetStats
Binarize

'output = tensor > threshold'.
Clip
EnsureDtype
FixedZeroMeanUnitVariance

normalize to zero mean, unit variance with precomputed values.
ScaleLinear
ScaleMeanVariance
ScaleRange
Sigmoid

1 / (1 + e^(-input)).
Softmax

Softmax activation function.
StardistPostprocessing2D
StardistPostprocessing3D
UpdateStats

Calculates sample and/or dataset measures
ZeroMeanUnitVariance

normalize to zero mean, unit variance.

Functions:

Name Description
get_proc

Attributes:

Name Type Description
NdBorder
NdTuple
ProcDescr
Processing

NdBorder module-attribute ¤

NdBorder = TypeVar('NdBorder', Tuple[Tuple[int, int], Tuple[int, int]], Tuple[Tuple[int, int], Tuple[int, int], Tuple[int, int]])
Used by:

NdTuple module-attribute ¤

NdTuple = TypeVar('NdTuple', Tuple[int, int], Tuple[int, int, int])
Used by:

ProcDescr module-attribute ¤

ProcDescr = Union[v0_4.PreprocessingDescr, v0_4.PostprocessingDescr, v0_5.PreprocessingDescr, v0_5.PostprocessingDescr]
Used by:

Processing module-attribute ¤

Processing = Union[AddKnownDatasetStats, Binarize, Clip, EnsureDtype, FixedZeroMeanUnitVariance, ScaleLinear, ScaleMeanVariance, ScaleRange, Sigmoid, StardistPostprocessing2D, StardistPostprocessing3D, Softmax, UpdateStats, ZeroMeanUnitVariance]
Returned by: Used by:

AddKnownDatasetStats dataclass ¤

AddKnownDatasetStats(dataset_stats: Mapping[DatasetMeasure, MeasureValue])

Bases: BlockwiseOperator


              flowchart TD
              bioimageio.core.proc_ops.AddKnownDatasetStats[AddKnownDatasetStats]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core.proc_ops.AddKnownDatasetStats
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                



              click bioimageio.core.proc_ops.AddKnownDatasetStats href "" "bioimageio.core.proc_ops.AddKnownDatasetStats"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__

Attributes:

Name Type Description
dataset_stats Mapping[DatasetMeasure, MeasureValue]
required_measures Collection[Measure]

dataset_stats instance-attribute ¤

dataset_stats: Mapping[DatasetMeasure, MeasureValue]

required_measures property ¤

required_measures: Collection[Measure]

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/proc_ops.py 
79
80
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    sample.stat.update(self.dataset_stats.items())

Binarize dataclass ¤

Binarize(input: MemberId, output: MemberId, threshold: Union[float, Sequence[float]], axis: Optional[AxisId] = None)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.Binarize[Binarize]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.Binarize
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.Binarize href "" "bioimageio.core.proc_ops.Binarize"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"

'output = tensor > threshold'.

Used by:

Methods:

Name Description
__call__
from_proc_descr
get_output_shape

Attributes:

Name Type Description
axis Optional[AxisId]
input MemberId
output MemberId
required_measures Collection[Measure]
threshold Union[float, Sequence[float]]

axis class-attribute instance-attribute ¤

axis: Optional[AxisId] = None

input instance-attribute ¤

input: MemberId

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures: Collection[Measure]

threshold instance-attribute ¤

threshold: Union[float, Sequence[float]]

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

from_proc_descr classmethod ¤

from_proc_descr(descr: Union[v0_4.BinarizeDescr, v0_5.BinarizeDescr], member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
@classmethod
def from_proc_descr(
    cls, descr: Union[v0_4.BinarizeDescr, v0_5.BinarizeDescr], member_id: MemberId
) -> Self:
    if isinstance(descr.kwargs, (v0_4.BinarizeKwargs, v0_5.BinarizeKwargs)):
        return cls(
            input=member_id, output=member_id, threshold=descr.kwargs.threshold
        )
    elif isinstance(descr.kwargs, v0_5.BinarizeAlongAxisKwargs):
        return cls(
            input=member_id,
            output=member_id,
            threshold=descr.kwargs.threshold,
            axis=descr.kwargs.axis,
        )
    else:
        assert_never(descr.kwargs)

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
163
164
165
166
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

Clip dataclass ¤

Clip(input: MemberId, output: MemberId, min: Optional[Union[float, SampleQuantile, DatasetQuantile]] = None, max: Optional[Union[float, SampleQuantile, DatasetQuantile]] = None)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.Clip[Clip]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.Clip
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.Clip href "" "bioimageio.core.proc_ops.Clip"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__
__post_init__
from_proc_descr
get_output_shape

Attributes:

Name Type Description
input MemberId
max Optional[Union[float, SampleQuantile, DatasetQuantile]]

maximum value for clipping
min Optional[Union[float, SampleQuantile, DatasetQuantile]]

minimum value for clipping
output MemberId
required_measures

input instance-attribute ¤

input: MemberId

max class-attribute instance-attribute ¤

max: Optional[Union[float, SampleQuantile, DatasetQuantile]] = None

maximum value for clipping

min class-attribute instance-attribute ¤

min: Optional[Union[float, SampleQuantile, DatasetQuantile]] = None

minimum value for clipping

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

__post_init__ ¤

__post_init__()
Source code in src/bioimageio/core/proc_ops.py 
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
def __post_init__(self):
    if self.min is None and self.max is None:
        raise ValueError("missing min or max value")

    if (
        isinstance(self.min, float)
        and isinstance(self.max, float)
        and self.min >= self.max
    ):
        raise ValueError(f"expected min < max, but {self.min} >= {self.max}")

    if isinstance(self.min, (SampleQuantile, DatasetQuantile)) and isinstance(
        self.max, (SampleQuantile, DatasetQuantile)
    ):
        if self.min.axes != self.max.axes:
            raise NotImplementedError(
                f"expected min and max quantiles with same axes, but got {self.min.axes} and {self.max.axes}"
            )
        if self.min.q >= self.max.q:
            raise ValueError(
                f"expected min quantile < max quantile, but {self.min.q} >= {self.max.q}"
            )

from_proc_descr classmethod ¤

from_proc_descr(descr: Union[v0_4.ClipDescr, v0_5.ClipDescr], member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
@classmethod
def from_proc_descr(
    cls, descr: Union[v0_4.ClipDescr, v0_5.ClipDescr], member_id: MemberId
) -> Self:
    if isinstance(descr, v0_5.ClipDescr):
        dataset_mode, axes = _get_axes(descr.kwargs)
        if dataset_mode:
            Quantile = DatasetQuantile
        else:
            Quantile = partial(SampleQuantile, method="inverted_cdf")

        if descr.kwargs.min is not None:
            min_arg = descr.kwargs.min
        elif descr.kwargs.min_percentile is not None:
            min_arg = Quantile(
                q=descr.kwargs.min_percentile / 100,
                axes=axes,
                member_id=member_id,
            )
        else:
            min_arg = None

        if descr.kwargs.max is not None:
            max_arg = descr.kwargs.max
        elif descr.kwargs.max_percentile is not None:
            max_arg = Quantile(
                q=descr.kwargs.max_percentile / 100,
                axes=axes,
                member_id=member_id,
            )
        else:
            max_arg = None

    elif isinstance(descr, v0_4.ClipDescr):
        min_arg = descr.kwargs.min
        max_arg = descr.kwargs.max
    else:
        assert_never(descr)

    return cls(
        input=member_id,
        output=member_id,
        min=min_arg,
        max=max_arg,
    )

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
259
260
261
262
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

EnsureDtype dataclass ¤

EnsureDtype(input: MemberId, output: MemberId, dtype: DTypeStr)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.EnsureDtype[EnsureDtype]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.EnsureDtype
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.EnsureDtype href "" "bioimageio.core.proc_ops.EnsureDtype"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__
from_proc_descr
get_descr
get_output_shape

Attributes:

Name Type Description
dtype DTypeStr
input MemberId
output MemberId
required_measures Collection[Measure]

dtype instance-attribute ¤

dtype: DTypeStr

input instance-attribute ¤

input: MemberId

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures: Collection[Measure]

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

from_proc_descr classmethod ¤

from_proc_descr(descr: v0_5.EnsureDtypeDescr, member_id: MemberId)
Source code in src/bioimageio/core/proc_ops.py 
315
316
317
@classmethod
def from_proc_descr(cls, descr: v0_5.EnsureDtypeDescr, member_id: MemberId):
    return cls(input=member_id, output=member_id, dtype=descr.kwargs.dtype)

get_descr ¤

get_descr()
Source code in src/bioimageio/core/proc_ops.py 
319
320
def get_descr(self):
    return v0_5.EnsureDtypeDescr(kwargs=v0_5.EnsureDtypeKwargs(dtype=self.dtype))

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
322
323
324
325
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

FixedZeroMeanUnitVariance dataclass ¤

FixedZeroMeanUnitVariance(input: MemberId, output: MemberId, mean: Union[float, xr.DataArray], std: Union[float, xr.DataArray], eps: float = 1e-06)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.FixedZeroMeanUnitVariance[FixedZeroMeanUnitVariance]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.FixedZeroMeanUnitVariance
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.FixedZeroMeanUnitVariance href "" "bioimageio.core.proc_ops.FixedZeroMeanUnitVariance"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"

normalize to zero mean, unit variance with precomputed values.

Used by:

Methods:

Name Description
__call__
__post_init__
from_proc_descr
get_descr
get_output_shape

Attributes:

Name Type Description
eps float
input MemberId
mean Union[float, xr.DataArray]
output MemberId
required_measures Collection[Measure]
std Union[float, xr.DataArray]

eps class-attribute instance-attribute ¤

eps: float = 1e-06

input instance-attribute ¤

input: MemberId

mean instance-attribute ¤

mean: Union[float, xr.DataArray]

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures: Collection[Measure]

std instance-attribute ¤

std: Union[float, xr.DataArray]

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

__post_init__ ¤

__post_init__()
Source code in src/bioimageio/core/proc_ops.py 
897
898
899
900
901
902
def __post_init__(self):
    assert (
        isinstance(self.mean, (int, float))
        or isinstance(self.std, (int, float))
        or self.mean.dims == self.std.dims
    )

from_proc_descr classmethod ¤

from_proc_descr(descr: v0_5.FixedZeroMeanUnitVarianceDescr, member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
@classmethod
def from_proc_descr(
    cls,
    descr: v0_5.FixedZeroMeanUnitVarianceDescr,
    member_id: MemberId,
) -> Self:
    if isinstance(descr.kwargs, v0_5.FixedZeroMeanUnitVarianceKwargs):
        dims = None
    elif isinstance(descr.kwargs, v0_5.FixedZeroMeanUnitVarianceAlongAxisKwargs):
        dims = (AxisId(descr.kwargs.axis),)
    else:
        assert_never(descr.kwargs)

    return cls(
        input=member_id,
        output=member_id,
        mean=xr.DataArray(descr.kwargs.mean, dims=dims),
        std=xr.DataArray(descr.kwargs.std, dims=dims),
    )

get_descr ¤

get_descr()
Source code in src/bioimageio/core/proc_ops.py 
933
934
935
936
937
938
939
940
941
942
943
944
945
946
def get_descr(self):
    if isinstance(self.mean, (int, float)):
        assert isinstance(self.std, (int, float))
        kwargs = v0_5.FixedZeroMeanUnitVarianceKwargs(mean=self.mean, std=self.std)
    else:
        assert isinstance(self.std, xr.DataArray)
        assert len(self.mean.dims) == 1
        kwargs = v0_5.FixedZeroMeanUnitVarianceAlongAxisKwargs(
            axis=AxisId(str(self.mean.dims[0])),
            mean=list(self.mean),
            std=list(self.std),
        )

    return v0_5.FixedZeroMeanUnitVarianceDescr(kwargs=kwargs)

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
908
909
910
911
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

ScaleLinear dataclass ¤

ScaleLinear(input: MemberId, output: MemberId, gain: Union[float, xr.DataArray] = 1.0, offset: Union[float, xr.DataArray] = 0.0)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.ScaleLinear[ScaleLinear]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.ScaleLinear
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.ScaleLinear href "" "bioimageio.core.proc_ops.ScaleLinear"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__
from_proc_descr
get_output_shape

Attributes:

Name Type Description
gain Union[float, xr.DataArray]

multiplicative factor
input MemberId
offset Union[float, xr.DataArray]

additive term
output MemberId
required_measures Collection[Measure]

gain class-attribute instance-attribute ¤

gain: Union[float, xr.DataArray] = 1.0

multiplicative factor

input instance-attribute ¤

input: MemberId

offset class-attribute instance-attribute ¤

offset: Union[float, xr.DataArray] = 0.0

additive term

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures: Collection[Measure]

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

from_proc_descr classmethod ¤

from_proc_descr(descr: Union[v0_4.ScaleLinearDescr, v0_5.ScaleLinearDescr], member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
@classmethod
def from_proc_descr(
    cls,
    descr: Union[v0_4.ScaleLinearDescr, v0_5.ScaleLinearDescr],
    member_id: MemberId,
) -> Self:
    kwargs = descr.kwargs
    if isinstance(kwargs, v0_5.ScaleLinearKwargs):
        axis = None
    elif isinstance(kwargs, v0_5.ScaleLinearAlongAxisKwargs):
        axis = kwargs.axis
    elif isinstance(kwargs, v0_4.ScaleLinearKwargs):
        if kwargs.axes is not None:
            raise NotImplementedError(
                "model.v0_4.ScaleLinearKwargs with axes not implemented, please consider updating the model to v0_5."
            )
        axis = None
    else:
        assert_never(kwargs)

    if axis:
        gain = xr.DataArray(np.atleast_1d(kwargs.gain), dims=axis)
        offset = xr.DataArray(np.atleast_1d(kwargs.offset), dims=axis)
    else:
        assert isinstance(kwargs.gain, (float, int)) or len(kwargs.gain) == 1, (
            kwargs.gain
        )
        gain = (
            kwargs.gain if isinstance(kwargs.gain, (float, int)) else kwargs.gain[0]
        )
        assert isinstance(kwargs.offset, (float, int)) or len(kwargs.offset) == 1
        offset = (
            kwargs.offset
            if isinstance(kwargs.offset, (float, int))
            else kwargs.offset[0]
        )

    return cls(input=member_id, output=member_id, gain=gain, offset=offset)

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
350
351
352
353
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

ScaleMeanVariance dataclass ¤

ScaleMeanVariance(input: MemberId, output: MemberId, axes: Optional[Sequence[AxisId]] = None, reference_tensor: Optional[MemberId] = None, eps: float = 1e-06)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.ScaleMeanVariance[ScaleMeanVariance]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.ScaleMeanVariance
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.ScaleMeanVariance href "" "bioimageio.core.proc_ops.ScaleMeanVariance"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__
__post_init__
from_proc_descr
get_output_shape

Attributes:

Name Type Description
axes Optional[Sequence[AxisId]]
eps float
input MemberId
mean Union[SampleMean, DatasetMean]
output MemberId
ref_mean Union[SampleMean, DatasetMean]
ref_std Union[SampleStd, DatasetStd]
reference_tensor Optional[MemberId]
required_measures
std Union[SampleStd, DatasetStd]

axes class-attribute instance-attribute ¤

axes: Optional[Sequence[AxisId]] = None

eps class-attribute instance-attribute ¤

eps: float = 1e-06

input instance-attribute ¤

input: MemberId

mean class-attribute instance-attribute ¤

mean: Union[SampleMean, DatasetMean] = field(init=False)

output instance-attribute ¤

output: MemberId

ref_mean class-attribute instance-attribute ¤

ref_mean: Union[SampleMean, DatasetMean] = field(init=False)

ref_std class-attribute instance-attribute ¤

ref_std: Union[SampleStd, DatasetStd] = field(init=False)

reference_tensor class-attribute instance-attribute ¤

reference_tensor: Optional[MemberId] = None

required_measures property ¤

required_measures

std class-attribute instance-attribute ¤

std: Union[SampleStd, DatasetStd] = field(init=False)

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

__post_init__ ¤

__post_init__()
Source code in src/bioimageio/core/proc_ops.py 
409
410
411
412
413
414
415
416
417
418
419
420
421
422
def __post_init__(self):
    axes = None if self.axes is None else tuple(self.axes)
    ref_tensor = self.reference_tensor or self.input
    if axes is None or AxisId("batch") not in axes:
        Mean = SampleMean
        Std = SampleStd
    else:
        Mean = DatasetMean
        Std = DatasetStd

    self.mean = Mean(member_id=self.input, axes=axes)
    self.std = Std(member_id=self.input, axes=axes)
    self.ref_mean = Mean(member_id=ref_tensor, axes=axes)
    self.ref_std = Std(member_id=ref_tensor, axes=axes)

from_proc_descr classmethod ¤

from_proc_descr(descr: Union[v0_4.ScaleMeanVarianceDescr, v0_5.ScaleMeanVarianceDescr], member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
@classmethod
def from_proc_descr(
    cls,
    descr: Union[v0_4.ScaleMeanVarianceDescr, v0_5.ScaleMeanVarianceDescr],
    member_id: MemberId,
) -> Self:
    kwargs = descr.kwargs
    _, axes = _get_axes(descr.kwargs)

    return cls(
        input=member_id,
        output=member_id,
        reference_tensor=MemberId(str(kwargs.reference_tensor)),
        axes=axes,
        eps=kwargs.eps,
    )

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
431
432
433
434
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

ScaleRange dataclass ¤

ScaleRange(input: MemberId, output: MemberId, lower_quantile: InitVar[Optional[Union[SampleQuantile, DatasetQuantile]]] = None, upper_quantile: InitVar[Optional[Union[SampleQuantile, DatasetQuantile]]] = None, eps: float = 1e-06)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.ScaleRange[ScaleRange]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.ScaleRange
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.ScaleRange href "" "bioimageio.core.proc_ops.ScaleRange"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__
__post_init__
from_proc_descr
get_descr
get_output_shape

Attributes:

Name Type Description
eps float
input MemberId
lower Union[SampleQuantile, DatasetQuantile]
output MemberId
required_measures
upper Union[SampleQuantile, DatasetQuantile]

eps class-attribute instance-attribute ¤

eps: float = 1e-06

input instance-attribute ¤

input: MemberId

lower class-attribute instance-attribute ¤

lower: Union[SampleQuantile, DatasetQuantile] = field(init=False)

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures

upper class-attribute instance-attribute ¤

upper: Union[SampleQuantile, DatasetQuantile] = field(init=False)

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

__post_init__ ¤

__post_init__(lower_quantile: Optional[Union[SampleQuantile, DatasetQuantile]], upper_quantile: Optional[Union[SampleQuantile, DatasetQuantile]])
Source code in src/bioimageio/core/proc_ops.py 
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
def __post_init__(
    self,
    lower_quantile: Optional[Union[SampleQuantile, DatasetQuantile]],
    upper_quantile: Optional[Union[SampleQuantile, DatasetQuantile]],
):
    if lower_quantile is None:
        tid = self.input if upper_quantile is None else upper_quantile.member_id
        self.lower = DatasetQuantile(q=0.0, member_id=tid)
    else:
        self.lower = lower_quantile

    if upper_quantile is None:
        self.upper = DatasetQuantile(q=1.0, member_id=self.lower.member_id)
    else:
        self.upper = upper_quantile

    assert self.lower.member_id == self.upper.member_id
    assert self.lower.q < self.upper.q
    assert self.lower.axes == self.upper.axes

from_proc_descr classmethod ¤

from_proc_descr(descr: Union[v0_4.ScaleRangeDescr, v0_5.ScaleRangeDescr], member_id: MemberId)
Source code in src/bioimageio/core/proc_ops.py 
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
@classmethod
def from_proc_descr(
    cls,
    descr: Union[v0_4.ScaleRangeDescr, v0_5.ScaleRangeDescr],
    member_id: MemberId,
):
    kwargs = descr.kwargs
    ref_tensor = (
        member_id
        if kwargs.reference_tensor is None
        else MemberId(str(kwargs.reference_tensor))
    )
    dataset_mode, axes = _get_axes(descr.kwargs)
    if dataset_mode:
        Quantile = DatasetQuantile
    else:
        Quantile = partial(SampleQuantile, method="linear")

    return cls(
        input=member_id,
        output=member_id,
        lower_quantile=Quantile(
            q=kwargs.min_percentile / 100,
            axes=axes,
            member_id=ref_tensor,
        ),
        upper_quantile=Quantile(
            q=kwargs.max_percentile / 100,
            axes=axes,
            member_id=ref_tensor,
        ),
    )

get_descr ¤

get_descr()
Source code in src/bioimageio/core/proc_ops.py 
553
554
555
556
557
558
559
560
561
562
563
564
565
def get_descr(self):
    assert self.lower.axes == self.upper.axes
    assert self.lower.member_id == self.upper.member_id

    return v0_5.ScaleRangeDescr(
        kwargs=v0_5.ScaleRangeKwargs(
            axes=self.lower.axes,
            min_percentile=self.lower.q * 100,
            max_percentile=self.upper.q * 100,
            eps=self.eps,
            reference_tensor=self.lower.member_id,
        )
    )

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
510
511
512
513
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

Sigmoid dataclass ¤

Sigmoid(input: MemberId, output: MemberId)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.Sigmoid[Sigmoid]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.Sigmoid
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.Sigmoid href "" "bioimageio.core.proc_ops.Sigmoid"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"

1 / (1 + e^(-input)).

Used by:

Methods:

Name Description
__call__
from_proc_descr
get_descr
get_output_shape

Attributes:

Name Type Description
input MemberId
output MemberId
required_measures Collection[Measure]

input instance-attribute ¤

input: MemberId

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures: Collection[Measure]

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

from_proc_descr classmethod ¤

from_proc_descr(descr: Union[v0_4.SigmoidDescr, v0_5.SigmoidDescr], member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
584
585
586
587
588
589
@classmethod
def from_proc_descr(
    cls, descr: Union[v0_4.SigmoidDescr, v0_5.SigmoidDescr], member_id: MemberId
) -> Self:
    assert isinstance(descr, (v0_4.SigmoidDescr, v0_5.SigmoidDescr))
    return cls(input=member_id, output=member_id)

get_descr ¤

get_descr()
Source code in src/bioimageio/core/proc_ops.py 
591
592
def get_descr(self):
    return v0_5.SigmoidDescr()

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
579
580
581
582
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

Softmax dataclass ¤

Softmax(input: MemberId, output: MemberId, axis: AxisId = AxisId('channel'))

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.Softmax[Softmax]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.Softmax
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.Softmax href "" "bioimageio.core.proc_ops.Softmax"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"

Softmax activation function.

Used by:

Methods:

Name Description
__call__
from_proc_descr
get_descr
get_output_shape

Attributes:

Name Type Description
axis AxisId
input MemberId
output MemberId
required_measures Collection[Measure]

axis class-attribute instance-attribute ¤

axis: AxisId = AxisId('channel')

input instance-attribute ¤

input: MemberId

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures: Collection[Measure]

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

from_proc_descr classmethod ¤

from_proc_descr(descr: v0_5.SoftmaxDescr, member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
616
617
618
619
@classmethod
def from_proc_descr(cls, descr: v0_5.SoftmaxDescr, member_id: MemberId) -> Self:
    assert isinstance(descr, v0_5.SoftmaxDescr)
    return cls(input=member_id, output=member_id, axis=descr.kwargs.axis)

get_descr ¤

get_descr()
Source code in src/bioimageio/core/proc_ops.py 
621
622
def get_descr(self):
    return v0_5.SoftmaxDescr(kwargs=v0_5.SoftmaxKwargs(axis=self.axis))

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
611
612
613
614
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

StardistPostprocessing2D dataclass ¤

StardistPostprocessing2D(prob_dist_input_id: MemberId, instance_labels_output_id: MemberId, grid: NdTuple, prob_threshold: float, nms_threshold: float, b: Union[int, NdBorder])

Bases: _StardistPostprocessingBase[Tuple[int, int], Tuple[Tuple[int, int], Tuple[int, int]]]


              flowchart TD
              bioimageio.core.proc_ops.StardistPostprocessing2D[StardistPostprocessing2D]
              bioimageio.core.proc_ops._StardistPostprocessingBase[_StardistPostprocessingBase]
              bioimageio.core._op_base.SamplewiseOperator[SamplewiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core.proc_ops._StardistPostprocessingBase --> bioimageio.core.proc_ops.StardistPostprocessing2D
                                bioimageio.core._op_base.SamplewiseOperator --> bioimageio.core.proc_ops._StardistPostprocessingBase
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.SamplewiseOperator
                




              click bioimageio.core.proc_ops.StardistPostprocessing2D href "" "bioimageio.core.proc_ops.StardistPostprocessing2D"
              click bioimageio.core.proc_ops._StardistPostprocessingBase href "" "bioimageio.core.proc_ops._StardistPostprocessingBase"
              click bioimageio.core._op_base.SamplewiseOperator href "" "bioimageio.core._op_base.SamplewiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__
from_proc_descr

Attributes:

Name Type Description
b Union[int, NdBorder]

Border region in which object probability is set to zero.
grid NdTuple

Grid size of network predictions.
instance_labels_output_id MemberId
nms_threshold float

The IoU threshold for non-maximum suppression.
prob_dist_input_id MemberId
prob_threshold float

Object probability threshold for non-maximum suppression.
required_measures Collection[Measure]

b instance-attribute ¤

b: Union[int, NdBorder]

Border region in which object probability is set to zero.

grid instance-attribute ¤

grid: NdTuple

Grid size of network predictions.

instance_labels_output_id instance-attribute ¤

instance_labels_output_id: MemberId

nms_threshold instance-attribute ¤

nms_threshold: float

The IoU threshold for non-maximum suppression.

prob_dist_input_id instance-attribute ¤

prob_dist_input_id: MemberId

prob_threshold instance-attribute ¤

prob_threshold: float

Object probability threshold for non-maximum suppression.

required_measures property ¤

required_measures: Collection[Measure]

__call__ ¤

__call__(sample: Sample) -> None
Source code in src/bioimageio/core/proc_ops.py 
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
def __call__(self, sample: Sample) -> None:
    prob_dist = sample.members[self.prob_dist_input_id]

    assert AxisId("channel") in prob_dist.dims, (
        "expected 'channel' axis in stardist probability/distance input"
    )
    allowed_spatial = tuple(
        map(AxisId, ("y", "x") if len(self.grid) == 2 else ("z", "y", "x"))
    )
    assert all(
        a in allowed_spatial or a in (AxisId("batch"), AxisId("channel"))
        for a in prob_dist.dims
    ), (
        f"expected prob_dist to have only 'batch', 'channel', and spatial axes {allowed_spatial}, but got {prob_dist.dims}"
    )

    spatial_shape = tuple(
        prob_dist.tagged_shape[a] * g for a, g in zip(allowed_spatial, self.grid)
    )
    if len(spatial_shape) != len(self.grid):
        raise ValueError(
            f"expected {len(self.grid)} spatial dimensions in prob_dist tensor, but got {len(spatial_shape)}"
        )
    else:
        spatial_shape = cast(NdTuple, spatial_shape)

    prob_dist = prob_dist.transpose(
        (AxisId("batch"), *allowed_spatial, AxisId("channel"))
    )
    labels: List[NDArray[Any]] = []
    for batch_idx in range(prob_dist.sizes[AxisId("batch")]):
        prob = prob_dist[
            {AxisId("batch"): batch_idx, AxisId("channel"): 0}
        ].to_numpy()
        dist = prob_dist[
            {AxisId("batch"): batch_idx, AxisId("channel"): slice(1, None)}
        ].to_numpy()

        labels_i = self._impl(prob, dist, spatial_shape)
        assert labels_i.shape == spatial_shape, (
            f"expected label image shape {spatial_shape}, but got {labels_i.shape}"
        )
        labels.append(labels_i)

    instance_labels = Tensor(
        np.stack(labels)[..., None],
        dims=(AxisId("batch"), *allowed_spatial, AxisId("channel")),
    )
    sample.members[self.instance_labels_output_id] = instance_labels

from_proc_descr classmethod ¤

from_proc_descr(descr: v0_5.StardistPostprocessingDescr, member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
@classmethod
def from_proc_descr(
    cls, descr: v0_5.StardistPostprocessingDescr, member_id: MemberId
) -> Self:
    if not isinstance(descr.kwargs, v0_5.StardistPostprocessingKwargs2D):
        raise TypeError(
            f"expected v0_5.StardistPostprocessingKwargs2D for 2D stardist post-processing, but got {type(descr.kwargs)}"
        )

    kwargs = descr.kwargs
    return cls(
        prob_dist_input_id=member_id,
        instance_labels_output_id=member_id,
        grid=kwargs.grid,
        prob_threshold=kwargs.prob_threshold,
        nms_threshold=kwargs.nms_threshold,
        b=kwargs.b,
    )

StardistPostprocessing3D dataclass ¤

StardistPostprocessing3D(prob_dist_input_id: MemberId, instance_labels_output_id: MemberId, grid: NdTuple, prob_threshold: float, nms_threshold: float, b: Union[int, NdBorder], n_rays: int, anisotropy: Tuple[float, float, float], overlap_label: Optional[int] = None)

Bases: _StardistPostprocessingBase[Tuple[int, int, int], Tuple[Tuple[int, int], Tuple[int, int], Tuple[int, int]]]


              flowchart TD
              bioimageio.core.proc_ops.StardistPostprocessing3D[StardistPostprocessing3D]
              bioimageio.core.proc_ops._StardistPostprocessingBase[_StardistPostprocessingBase]
              bioimageio.core._op_base.SamplewiseOperator[SamplewiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core.proc_ops._StardistPostprocessingBase --> bioimageio.core.proc_ops.StardistPostprocessing3D
                                bioimageio.core._op_base.SamplewiseOperator --> bioimageio.core.proc_ops._StardistPostprocessingBase
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.SamplewiseOperator
                




              click bioimageio.core.proc_ops.StardistPostprocessing3D href "" "bioimageio.core.proc_ops.StardistPostprocessing3D"
              click bioimageio.core.proc_ops._StardistPostprocessingBase href "" "bioimageio.core.proc_ops._StardistPostprocessingBase"
              click bioimageio.core._op_base.SamplewiseOperator href "" "bioimageio.core._op_base.SamplewiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"
Used by:

Methods:

Name Description
__call__
from_proc_descr

Attributes:

Name Type Description
anisotropy Tuple[float, float, float]

Anisotropy factors for 3D star-convex polyhedra, i.e. the physical pixel size along each spatial axis.
b Union[int, NdBorder]

Border region in which object probability is set to zero.
grid NdTuple

Grid size of network predictions.
instance_labels_output_id MemberId
n_rays int

Number of rays for 3D star-convex polyhedra.
nms_threshold float

The IoU threshold for non-maximum suppression.
overlap_label Optional[int]

Optional label to apply to any area of overlapping predicted objects.
prob_dist_input_id MemberId
prob_threshold float

Object probability threshold for non-maximum suppression.
required_measures Collection[Measure]

anisotropy instance-attribute ¤

anisotropy: Tuple[float, float, float]

Anisotropy factors for 3D star-convex polyhedra, i.e. the physical pixel size along each spatial axis.

b instance-attribute ¤

b: Union[int, NdBorder]

Border region in which object probability is set to zero.

grid instance-attribute ¤

grid: NdTuple

Grid size of network predictions.

instance_labels_output_id instance-attribute ¤

instance_labels_output_id: MemberId

n_rays instance-attribute ¤

n_rays: int

Number of rays for 3D star-convex polyhedra.

nms_threshold instance-attribute ¤

nms_threshold: float

The IoU threshold for non-maximum suppression.

overlap_label class-attribute instance-attribute ¤

overlap_label: Optional[int] = None

Optional label to apply to any area of overlapping predicted objects.

prob_dist_input_id instance-attribute ¤

prob_dist_input_id: MemberId

prob_threshold instance-attribute ¤

prob_threshold: float

Object probability threshold for non-maximum suppression.

required_measures property ¤

required_measures: Collection[Measure]

__call__ ¤

__call__(sample: Sample) -> None
Source code in src/bioimageio/core/proc_ops.py 
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
def __call__(self, sample: Sample) -> None:
    prob_dist = sample.members[self.prob_dist_input_id]

    assert AxisId("channel") in prob_dist.dims, (
        "expected 'channel' axis in stardist probability/distance input"
    )
    allowed_spatial = tuple(
        map(AxisId, ("y", "x") if len(self.grid) == 2 else ("z", "y", "x"))
    )
    assert all(
        a in allowed_spatial or a in (AxisId("batch"), AxisId("channel"))
        for a in prob_dist.dims
    ), (
        f"expected prob_dist to have only 'batch', 'channel', and spatial axes {allowed_spatial}, but got {prob_dist.dims}"
    )

    spatial_shape = tuple(
        prob_dist.tagged_shape[a] * g for a, g in zip(allowed_spatial, self.grid)
    )
    if len(spatial_shape) != len(self.grid):
        raise ValueError(
            f"expected {len(self.grid)} spatial dimensions in prob_dist tensor, but got {len(spatial_shape)}"
        )
    else:
        spatial_shape = cast(NdTuple, spatial_shape)

    prob_dist = prob_dist.transpose(
        (AxisId("batch"), *allowed_spatial, AxisId("channel"))
    )
    labels: List[NDArray[Any]] = []
    for batch_idx in range(prob_dist.sizes[AxisId("batch")]):
        prob = prob_dist[
            {AxisId("batch"): batch_idx, AxisId("channel"): 0}
        ].to_numpy()
        dist = prob_dist[
            {AxisId("batch"): batch_idx, AxisId("channel"): slice(1, None)}
        ].to_numpy()

        labels_i = self._impl(prob, dist, spatial_shape)
        assert labels_i.shape == spatial_shape, (
            f"expected label image shape {spatial_shape}, but got {labels_i.shape}"
        )
        labels.append(labels_i)

    instance_labels = Tensor(
        np.stack(labels)[..., None],
        dims=(AxisId("batch"), *allowed_spatial, AxisId("channel")),
    )
    sample.members[self.instance_labels_output_id] = instance_labels

from_proc_descr classmethod ¤

from_proc_descr(descr: v0_5.StardistPostprocessingDescr, member_id: MemberId) -> Self
Source code in src/bioimageio/core/proc_ops.py 
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
@classmethod
def from_proc_descr(
    cls, descr: v0_5.StardistPostprocessingDescr, member_id: MemberId
) -> Self:
    if not isinstance(descr.kwargs, v0_5.StardistPostprocessingKwargs3D):
        raise TypeError(
            f"expected v0_5.StardistPostprocessingKwargs3D for 3D stardist post-processing, but got {type(descr.kwargs)}"
        )

    kwargs = descr.kwargs
    return cls(
        prob_dist_input_id=member_id,
        instance_labels_output_id=member_id,
        grid=kwargs.grid,
        prob_threshold=kwargs.prob_threshold,
        nms_threshold=kwargs.nms_threshold,
        n_rays=kwargs.n_rays,
        anisotropy=kwargs.anisotropy,
        b=kwargs.b,
        overlap_label=kwargs.overlap_label,
    )

UpdateStats dataclass ¤

UpdateStats(stats_calculator: StatsCalculator, keep_updating_initial_dataset_stats: bool = False)

Bases: SamplewiseOperator


              flowchart TD
              bioimageio.core.proc_ops.UpdateStats[UpdateStats]
              bioimageio.core._op_base.SamplewiseOperator[SamplewiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SamplewiseOperator --> bioimageio.core.proc_ops.UpdateStats
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.SamplewiseOperator
                



              click bioimageio.core.proc_ops.UpdateStats href "" "bioimageio.core.proc_ops.UpdateStats"
              click bioimageio.core._op_base.SamplewiseOperator href "" "bioimageio.core._op_base.SamplewiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"

Calculates sample and/or dataset measures

Used by:

Methods:

Name Description
__call__
__post_init__

Attributes:

Name Type Description
keep_updating_initial_dataset_stats bool

indicates if operator calls should keep updating initial dataset statistics or not;
required_measures Collection[Measure]
stats_calculator StatsCalculator

StatsCalculator to be used by this operator.

keep_updating_initial_dataset_stats class-attribute instance-attribute ¤

keep_updating_initial_dataset_stats: bool = False

indicates if operator calls should keep updating initial dataset statistics or not; if the stats_calculator was not provided with any initial dataset statistics, these are always updated with every new sample.

required_measures property ¤

required_measures: Collection[Measure]

stats_calculator instance-attribute ¤

stats_calculator: StatsCalculator

StatsCalculator to be used by this operator.

__call__ ¤

__call__(sample: Sample) -> None
Source code in src/bioimageio/core/proc_ops.py 
142
143
144
145
146
def __call__(self, sample: Sample) -> None:
    if self._keep_updating_dataset_stats:
        sample.stat.update(self.stats_calculator.update_and_get_all(sample))
    else:
        sample.stat.update(self.stats_calculator.skip_update_and_get_all(sample))

__post_init__ ¤

__post_init__()
Source code in src/bioimageio/core/proc_ops.py 
136
137
138
139
140
def __post_init__(self):
    self._keep_updating_dataset_stats = (
        self.keep_updating_initial_dataset_stats
        or not self.stats_calculator.has_dataset_measures
    )

ZeroMeanUnitVariance dataclass ¤

ZeroMeanUnitVariance(input: MemberId, output: MemberId, mean: MeanMeasure, std: StdMeasure, eps: float = 1e-06)

Bases: SimpleOperator


              flowchart TD
              bioimageio.core.proc_ops.ZeroMeanUnitVariance[ZeroMeanUnitVariance]
              bioimageio.core._op_base.SimpleOperator[SimpleOperator]
              bioimageio.core._op_base.BlockwiseOperator[BlockwiseOperator]
              bioimageio.core._op_base.Operator[Operator]

                              bioimageio.core._op_base.SimpleOperator --> bioimageio.core.proc_ops.ZeroMeanUnitVariance
                                bioimageio.core._op_base.BlockwiseOperator --> bioimageio.core._op_base.SimpleOperator
                                bioimageio.core._op_base.Operator --> bioimageio.core._op_base.BlockwiseOperator
                




              click bioimageio.core.proc_ops.ZeroMeanUnitVariance href "" "bioimageio.core.proc_ops.ZeroMeanUnitVariance"
              click bioimageio.core._op_base.SimpleOperator href "" "bioimageio.core._op_base.SimpleOperator"
              click bioimageio.core._op_base.BlockwiseOperator href "" "bioimageio.core._op_base.BlockwiseOperator"
              click bioimageio.core._op_base.Operator href "" "bioimageio.core._op_base.Operator"

normalize to zero mean, unit variance.

Used by:

Methods:

Name Description
__call__
__post_init__
from_proc_descr
get_descr
get_output_shape

Attributes:

Name Type Description
eps float
input MemberId
mean MeanMeasure
output MemberId
required_measures Set[Union[MeanMeasure, StdMeasure]]
std StdMeasure

eps class-attribute instance-attribute ¤

eps: float = 1e-06

input instance-attribute ¤

input: MemberId

mean instance-attribute ¤

mean: MeanMeasure

output instance-attribute ¤

output: MemberId

required_measures property ¤

required_measures: Set[Union[MeanMeasure, StdMeasure]]

std instance-attribute ¤

std: StdMeasure

__call__ ¤

__call__(sample: Union[Sample, SampleBlock]) -> None
Source code in src/bioimageio/core/_op_base.py 
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def __call__(self, sample: Union[Sample, SampleBlock]) -> None:
    if self.input not in sample.members:
        return

    input_tensor = sample.members[self.input]
    output_tensor = self._apply(input_tensor, sample.stat)

    if self.output in sample.members:
        assert (
            sample.members[self.output].tagged_shape == output_tensor.tagged_shape
        )

    if isinstance(sample, Sample):
        sample.members[self.output] = output_tensor
    elif isinstance(sample, SampleBlock):
        b = sample.blocks[self.input]
        sample.blocks[self.output] = Block(
            sample_shape=self.get_output_shape(sample.shape[self.input]),
            data=output_tensor,
            inner_slice=b.inner_slice,
            halo=b.halo,
            block_index=b.block_index,
            blocks_in_sample=b.blocks_in_sample,
        )
    else:
        assert_never(sample)

__post_init__ ¤

__post_init__()
Source code in src/bioimageio/core/proc_ops.py 
843
844
def __post_init__(self):
    assert self.mean.axes == self.std.axes

from_proc_descr classmethod ¤

from_proc_descr(descr: Union[v0_4.ZeroMeanUnitVarianceDescr, v0_5.ZeroMeanUnitVarianceDescr], member_id: MemberId)
Source code in src/bioimageio/core/proc_ops.py 
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
@classmethod
def from_proc_descr(
    cls,
    descr: Union[v0_4.ZeroMeanUnitVarianceDescr, v0_5.ZeroMeanUnitVarianceDescr],
    member_id: MemberId,
):
    dataset_mode, axes = _get_axes(descr.kwargs)

    if dataset_mode:
        Mean = DatasetMean
        Std = DatasetStd
    else:
        Mean = SampleMean
        Std = SampleStd

    return cls(
        input=member_id,
        output=member_id,
        mean=Mean(axes=axes, member_id=member_id),
        std=Std(axes=axes, member_id=member_id),
    )

get_descr ¤

get_descr()
Source code in src/bioimageio/core/proc_ops.py 
882
883
884
885
def get_descr(self):
    return v0_5.ZeroMeanUnitVarianceDescr(
        kwargs=v0_5.ZeroMeanUnitVarianceKwargs(axes=self.mean.axes, eps=self.eps)
    )

get_output_shape ¤

get_output_shape(input_shape: Mapping[AxisId, int]) -> Mapping[AxisId, int]
Source code in src/bioimageio/core/proc_ops.py 
850
851
852
853
def get_output_shape(
    self, input_shape: Mapping[AxisId, int]
) -> Mapping[AxisId, int]:
    return input_shape

get_proc ¤

get_proc(proc_descr: ProcDescr, tensor_descr: Union[v0_4.InputTensorDescr, v0_4.OutputTensorDescr, v0_5.InputTensorDescr, v0_5.OutputTensorDescr]) -> Processing
Source code in src/bioimageio/core/proc_ops.py 
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
def get_proc(
    proc_descr: ProcDescr,
    tensor_descr: Union[
        v0_4.InputTensorDescr,
        v0_4.OutputTensorDescr,
        v0_5.InputTensorDescr,
        v0_5.OutputTensorDescr,
    ],
) -> Processing:
    member_id = get_member_id(tensor_descr)

    if isinstance(proc_descr, (v0_4.BinarizeDescr, v0_5.BinarizeDescr)):
        return Binarize.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, (v0_4.ClipDescr, v0_5.ClipDescr)):
        return Clip.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, v0_5.EnsureDtypeDescr):
        return EnsureDtype.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, v0_5.FixedZeroMeanUnitVarianceDescr):
        return FixedZeroMeanUnitVariance.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, (v0_4.ScaleLinearDescr, v0_5.ScaleLinearDescr)):
        return ScaleLinear.from_proc_descr(proc_descr, member_id)
    elif isinstance(
        proc_descr, (v0_4.ScaleMeanVarianceDescr, v0_5.ScaleMeanVarianceDescr)
    ):
        return ScaleMeanVariance.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, (v0_4.ScaleRangeDescr, v0_5.ScaleRangeDescr)):
        return ScaleRange.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, (v0_4.SigmoidDescr, v0_5.SigmoidDescr)):
        return Sigmoid.from_proc_descr(proc_descr, member_id)
    elif (
        isinstance(proc_descr, v0_4.ZeroMeanUnitVarianceDescr)
        and proc_descr.kwargs.mode == "fixed"
    ):
        if not isinstance(
            tensor_descr, (v0_4.InputTensorDescr, v0_4.OutputTensorDescr)
        ):
            raise TypeError(
                "Expected v0_4 tensor description for v0_4 processing description"
            )

        v5_proc_descr = _convert_proc(proc_descr, tensor_descr.axes)
        assert isinstance(v5_proc_descr, v0_5.FixedZeroMeanUnitVarianceDescr)
        return FixedZeroMeanUnitVariance.from_proc_descr(v5_proc_descr, member_id)
    elif isinstance(
        proc_descr,
        (v0_4.ZeroMeanUnitVarianceDescr, v0_5.ZeroMeanUnitVarianceDescr),
    ):
        return ZeroMeanUnitVariance.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, v0_5.SoftmaxDescr):
        return Softmax.from_proc_descr(proc_descr, member_id)
    elif isinstance(proc_descr, v0_5.StardistPostprocessingDescr):
        if isinstance(proc_descr.kwargs, v0_5.StardistPostprocessingKwargs2D):
            return StardistPostprocessing2D.from_proc_descr(proc_descr, member_id)
        elif isinstance(proc_descr.kwargs, v0_5.StardistPostprocessingKwargs3D):
            return StardistPostprocessing3D.from_proc_descr(proc_descr, member_id)
        else:
            raise ValueError(
                f"expected ndim 2 or 3 for stardist postprocessing, but got {proc_descr.kwargs.ndim}"
            )
    else:
        assert_never(proc_descr)