π Geometric Computer Vision Library for Spatial AI
Python artificial-intelligencecomputer-visiondeep-learninghacktoberfestimage-processingmachine-learningneural-networkpythonpytorchroboticsspatial-ai
# LICENSE HEADER MANAGED BY add-license-header
#
# Copyright 2018 Kornia Team
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
import pytest
import torch
from torch._dynamo.testing import CompileCounter
from kornia.enhance.threshold import ThresholdType, threshold
from testing.base import DYNAMO_UNAVAILABLE_REASON, dynamo_is_available
class TestThreshold:
@pytest.mark.parametrize(
"ttype",
[
ThresholdType.THRESH_BINARY,
ThresholdType.THRESH_BINARY_INV,
ThresholdType.THRESH_TRUNC,
ThresholdType.THRESH_TOZERO,
ThresholdType.THRESH_TOZERO_INV,
],
)
@pytest.mark.parametrize("shape", [(1, 1, 5, 7), (2, 3, 11, 9)])
def test_output_properties(self, ttype, shape, device, dtype):
x = torch.rand(shape, device=device, dtype=dtype)
out = threshold(x, thresh=0.5, maxval=1.0, type=ttype)
assert out.shape == x.shape
assert out.dtype == x.dtype
assert out.device == x.device
def test_binary_rule_strict_greater(self, device, dtype):
x = torch.tensor([0.2, 0.5, 0.7], device=device, dtype=dtype)
out = threshold(x, thresh=0.5, maxval=9.0, type=ThresholdType.THRESH_BINARY)
expected = torch.tensor([0.0, 0.0, 9.0], device=device, dtype=dtype)
assert torch.allclose(out, expected)
def test_binary_inv_rule_strict_greater(self, device, dtype):
x = torch.tensor([0.2, 0.5, 0.7], device=device, dtype=dtype)
out = threshold(x, thresh=0.5, maxval=9.0, type=ThresholdType.THRESH_BINARY_INV)
expected = torch.tensor([9.0, 9.0, 0.0], device=device, dtype=dtype)
assert torch.allclose(out, expected)
def test_trunc(self, device, dtype):
x = torch.tensor([0.2, 0.5, 0.7], device=device, dtype=dtype)
out = threshold(x, thresh=0.5, maxval=9.0, type=ThresholdType.THRESH_TRUNC)
expected = torch.tensor([0.2, 0.5, 0.5], device=device, dtype=dtype)
assert torch.allclose(out, expected)
def test_tozero(self, device, dtype):
x = torch.tensor([0.2, 0.5, 0.7], device=device, dtype=dtype)
out = threshold(x, thresh=0.5, maxval=9.0, type=ThresholdType.THRESH_TOZERO)
expected = torch.tensor([0.0, 0.0, 0.7], device=device, dtype=dtype)
assert torch.allclose(out, expected)
def test_tozero_inv(self, device, dtype):
x = torch.tensor([0.2, 0.5, 0.7], device=device, dtype=dtype)
out = threshold(x, thresh=0.5, maxval=9.0, type=ThresholdType.THRESH_TOZERO_INV)
expected = torch.tensor([0.2, 0.5, 0.0], device=device, dtype=dtype)
assert torch.allclose(out, expected)
def test_otsu_raises(self, device, dtype):
x = torch.rand(1, 1, 5, 5, device=device, dtype=dtype)
with pytest.raises(NotImplementedError):
threshold(x, thresh=0.0, maxval=1.0, type=ThresholdType.THRESH_OTSU)
@pytest.mark.skipif(not dynamo_is_available(), reason=DYNAMO_UNAVAILABLE_REASON)
@pytest.mark.parametrize("ttype", [ThresholdType.THRESH_BINARY, ThresholdType.THRESH_TOZERO_INV])
def test_traces_in_one_graph_like_eager(self, ttype, device, dtype):
# The supported-type check must not iterate the enum class while tracing: torch 2.9's Dynamo cannot (#5819).
x = torch.rand(2, 3, 5, 7, device=device, dtype=dtype)
counter = CompileCounter()
torch._dynamo.reset()
traced = torch.compile(threshold, backend=counter, fullgraph=True)
torch.testing.assert_close(traced(x, 0.5, 1.0, ttype), threshold(x, 0.5, 1.0, ttype), rtol=0, atol=0)
assert counter.frame_count == 1
@pytest.mark.parametrize(
"image_dtype, values, thresh, ttype, expected",
[
# A threshold below the range passes every element instead of wrapping to 255 or raising.
(torch.uint8, [0, 100, 255], -1, ThresholdType.THRESH_BINARY, [9, 9, 9]),
(torch.uint8, [0, 100, 255], -0.5, ThresholdType.THRESH_BINARY, [9, 9, 9]),
(torch.uint8, [0, 100, 255], -1, ThresholdType.THRESH_TRUNC, [0, 0, 0]),
# A threshold above the range passes no element instead of raising.
(torch.uint8, [0, 100, 255], 300, ThresholdType.THRESH_BINARY_INV, [9, 9, 9]),
(torch.uint8, [0, 100, 255], 255.5, ThresholdType.THRESH_TOZERO, [0, 0, 0]),
# A negative fraction is rounded down, not truncated toward zero.
(torch.int16, [-2, -1, 0, 1], -0.5, ThresholdType.THRESH_BINARY, [0, 0, 9, 9]),
(torch.int16, [-2, -1, 0, 1], -1.5, ThresholdType.THRESH_TRUNC, [-2, -2, -2, -2]),
(torch.int16, [-2, -1, 0, 1], -1.5, ThresholdType.THRESH_TOZERO_INV, [-2, 0, 0, 0]),
(torch.uint8, [100, 101], 100.7, ThresholdType.THRESH_BINARY, [0, 9]),
(torch.int16, [-1, 0, 1], float("nan"), ThresholdType.THRESH_TOZERO, [0, 0, 0]),
(torch.int64, [2**60, 2**60 + 1, 2**60 + 2], 2**60 + 1, ThresholdType.THRESH_BINARY, [0, 0, 9]),
# A threshold at the dtype's minimum is in range: every element but the minimum passes.
(torch.uint8, [0, 1, 255], 0, ThresholdType.THRESH_BINARY, [0, 9, 9]),
# Rounded down below the range, a signed threshold truncates to the dtype's minimum, not to 0.
(torch.int8, [-128, 0, 127], -128.5, ThresholdType.THRESH_TRUNC, [-128, -128, -128]),
],
)
@pytest.mark.parametrize("tensor_thresh", [False, True])
def test_integer_input_compares_against_the_threshold_as_given(
self, image_dtype, values, thresh, ttype, expected, tensor_thresh, device
):
x = torch.tensor(values, device=device, dtype=image_dtype)
if tensor_thresh:
thresh = torch.tensor(thresh, device=device)
out = threshold(x, thresh=thresh, maxval=9, type=ttype)
assert out.dtype == image_dtype
assert out.tolist() == expected