def __call__(self, image, greyscale=False, image_diagonal=None,
group_prefix='dlib', n_upscales=0):
r"""
Perform a detection using the cached dlib detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
greyscale : `bool`, optional
Convert the image to greyscale or not.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
n_upscales : `int`, optional
Number of times to upscale the image when performing the detection,
may increase the chances of detecting smaller objects.
Returns
------
bounding_boxes : `list` of `menpo.shape.PointDirectedGraph`
The detected objects.
"""
detect_partial = partial(self._detector, n_upscales=n_upscales)
return detect(detect_partial, image, greyscale=greyscale,
image_diagonal=image_diagonal, group_prefix=group_prefix)
def __call__(self, image, greyscale=True, image_diagonal=None,
group_prefix='object', n_upscales=0):
r"""
Perform a detection using the cached dlib detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : menpo.image.Image
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
greyscale : bool, optional
Convert the image to greyscale or not.
image_diagonal : int, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : str, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
n_upscales : int, optional
Number of times to upscale the image when performing the detection,
may increase the chances of detecting smaller objects.
Returns
------
bounding_boxes : menpo.shape.PointDirectedGraph
The detected objects.
"""
detect_partial = partial(self._detector, n_upscales=n_upscales)
return detect(detect_partial, image, greyscale=greyscale,
image_diagonal=image_diagonal, group_prefix=group_prefix)
def __call__(self,
image,
greyscale=True,
image_diagonal=None,
group_prefix='ffld2',
padding=6,
interval=5,
threshold=0.5,
overlap=0.3):
r"""
Perform a detection using the cached ffdl2 detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
greyscale : `bool`, optional
Whether to convert the image to greyscale or not.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
padding : `int`, optional
Amount of zero padding in HOG cells
interval : `int`, optional
Number of levels per octave in the HOG pyramid
threshold : `double`, optional
Minimum detection threshold. Detections with a score less than this
value are not returned. Values can be negative.
overlap : `double`, optional
Minimum overlap in in latent positive search and
non-maxima suppression.
As discussed in the Face Detection Without Bells and Whistles paper,
a sensible value for overlap is 0.3
Returns
------
bounding_boxes : `menpo.shape.PointDirectedGraph`
The detected objects.
"""
detect_partial = partial(self._detector,
padding=padding,
interval=interval,
threshold=threshold,
overlap=overlap)
return detect(detect_partial,
image,
greyscale=greyscale,
image_diagonal=image_diagonal,
group_prefix=group_prefix)
def test_rescaling_image():
takeo_copy = takeo.copy()
ratio = 200.0 / takeo_copy.diagonal()
pcs = detect(fake_detector, takeo_copy, image_diagonal=200)
assert len(pcs) == 1
assert takeo_copy.n_channels == 3
assert takeo_copy.landmarks['object_0'][None].n_points == 4
assert_allclose(takeo_copy.landmarks['object_0'][None].points,
fake_box * (1.0 / ratio), atol=10e-2)
def test_rescaling_image():
takeo_copy = takeo.copy()
ratio = 200.0 / takeo_copy.diagonal()
pcs = detect(fake_detector, takeo_copy, image_diagonal=200)
assert len(pcs) == 1
assert takeo_copy.n_channels == 3
assert takeo_copy.landmarks['object_0'].n_points == 4
assert_allclose(takeo_copy.landmarks['object_0'].points,
fake_box * (1.0 / ratio),
atol=10e-2)
def __call__(self,
image,
image_diagonal=None,
group_prefix='opencv',
scale_factor=1.1,
min_neighbours=5,
min_size=(30, 30),
flags=cv2.cv.CV_HAAR_SCALE_IMAGE):
r"""
Perform a detection using the cached opencv detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
scale_factor : `float`, optional
The amount to increase the sliding windows by over the second
pass.
min_neighbours : `int`, optional
The minimum number of neighbours (close detections) before
Non-Maximum suppression to be considered a detection. Use 0
to return all detections.
min_size : `tuple` of 2 ints
The minimum object size in pixels that the detector will consider.
flags : `int`, optional
The flags to be passed through to the detector.
Returns
------
bounding_boxes : `menpo.shape.PointDirectedGraph`
The detected objects.
"""
detect_partial = partial(self._detector,
scale_factor=scale_factor,
min_neighbours=min_neighbours,
min_size=min_size,
flags=flags)
return detect(detect_partial,
image,
greyscale=True,
image_diagonal=image_diagonal,
group_prefix=group_prefix)
def __call__(
self,
image,
image_diagonal=None,
group_prefix="opencv",
scale_factor=1.1,
min_neighbours=5,
min_size=(30, 30),
flags=None,
):
r"""
Perform a detection using the cached opencv detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
scale_factor : `float`, optional
The amount to increase the sliding windows by over the second
pass.
min_neighbours : `int`, optional
The minimum number of neighbours (close detections) before
Non-Maximum suppression to be considered a detection. Use 0
to return all detections.
min_size : `tuple` of 2 ints
The minimum object size in pixels that the detector will consider.
flags : `int`, optional
The flags to be passed through to the detector.
Returns
------
bounding_boxes : `list` of `menpo.shape.PointDirectedGraph`
The detected objects.
"""
if flags is None:
flags = _get_default_flags()
detect_partial = partial(
self._detector, scale_factor=scale_factor, min_neighbours=min_neighbours, min_size=min_size, flags=flags
)
return detect(detect_partial, image, greyscale=True, image_diagonal=image_diagonal, group_prefix=group_prefix)
def __call__(self,
image,
greyscale=False,
image_diagonal=None,
group_prefix='bob',
threshold=20,
minimum_overlap=0.2):
r"""
Perform a detection using the cached bob detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
greyscale : `bool`, optional
Convert the image to greyscale or not.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
threshold : `float`, optional
The threshold of the quality of detected faces.
Detections with a quality lower than this value will not be
considered. Higher thresholds will not detect all faces, while lower
thresholds will generate false detections.
``minimum_overlap`` : `float` ``[0, 1]``
Computes the best detection using the given minimum overlap.
Returns
------
bounding_boxes : `list` of `menpo.shape.PointDirectedGraph`
The detected objects.
"""
detect_partial = partial(self._detector,
threshold=threshold,
minimum_overlap=minimum_overlap)
return detect(detect_partial,
image,
greyscale=greyscale,
image_diagonal=image_diagonal,
group_prefix=group_prefix,
channels_at_back=False)
def __call__(self, image, greyscale=True, image_diagonal=None,
group_prefix='ffld2', padding=6, interval=5, threshold=0.5,
overlap=0.3):
r"""
Perform a detection using the cached ffdl2 detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
greyscale : `bool`, optional
Whether to convert the image to greyscale or not.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
padding : `int`, optional
Amount of zero padding in HOG cells
interval : `int`, optional
Number of levels per octave in the HOG pyramid
threshold : `double`, optional
Minimum detection threshold. Detections with a score less than this
value are not returned. Values can be negative.
overlap : `double`, optional
Minimum overlap in in latent positive search and
non-maxima suppression.
As discussed in the Face Detection Without Bells and Whistles paper,
a sensible value for overlap is 0.3
Returns
------
bounding_boxes : `list` of `menpo.shape.PointDirectedGraph`
The detected objects.
"""
detect_partial = partial(self._detector, padding=padding,
interval=interval, threshold=threshold,
overlap=overlap)
return detect(detect_partial, image, greyscale=greyscale,
image_diagonal=image_diagonal, group_prefix=group_prefix)
def __call__(self, image, greyscale=False, image_diagonal=None,
group_prefix='bob', threshold=20, minimum_overlap=0.2):
r"""
Perform a detection using the cached bob detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
greyscale : `bool`, optional
Convert the image to greyscale or not.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
threshold : `float`, optional
The threshold of the quality of detected faces.
Detections with a quality lower than this value will not be
considered. Higher thresholds will not detect all faces, while lower
thresholds will generate false detections.
``minimum_overlap`` : `float` ``[0, 1]``
Computes the best detection using the given minimum overlap.
Returns
------
bounding_boxes : `list` of `menpo.shape.PointDirectedGraph`
The detected objects.
"""
detect_partial = partial(self._detector, threshold=threshold,
minimum_overlap=minimum_overlap)
return detect(detect_partial, image, greyscale=greyscale,
image_diagonal=image_diagonal, group_prefix=group_prefix,
channels_at_back=False)
def test_passing_uint8_rgb_image_greyscale_no_convert():
fake_image = MagicMock()
fake_image.n_channels = 3
pcs = detect(fake_detector, fake_image, greyscale=True)
assert len(pcs) == 1
fake_image.as_greyscale.assert_called_once_with(mode='luminosity')
def test_passing_uint8_image():
takeo_copy = takeo_uint8.copy()
pcs = detect(fake_detector, takeo_copy, greyscale=False)
assert len(pcs) == 1
assert takeo_copy.n_channels == 3
assert takeo_copy.landmarks['object_0'][None].n_points == 4
def __call__(self,
image,
image_diagonal=None,
group_prefix='object',
max_detections=100,
orientations=0.0,
scale_factor=1.2,
stride_factor=0.1,
min_size=100,
confidence_cutoff=3.0):
r"""
Perform a detection using the cached pico detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : menpo.image.Image
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
image_diagonal : int, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : str, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
max_detections : `int`, optional
The maximum number of detections to return.
orientations : list of `float`s or `float`, optional
The orientations of the cascades to use. ``0.0`` will perform an
axis aligned detection. Values greater than ``0.0`` will perform
detections of the cascade rotated counterclockwise around a unit
circle.
If a list is passed, each item should be an orientation in
radians around the unit circle, with ``0.0`` being axis aligned.
scale_factor : `float`, optional
The ratio to increase the cascade window at every iteration. Must
be greater than 1.0
stride_factor : `float`, optional
The ratio to decrease the window step by at every iteration. Must be
less than 1.0, optional
min_size : `float`, optional
The minimum size in pixels (diameter of the detection circle) that a
face can be. This is the starting cascade window size.
confidence_cutoff : `float`, optional
The confidence value to trim the detections with. Any detections
with confidence less than the cutoff will be discarded.
Returns
------
bounding_boxes : menpo.shape.PointDirectedGraph
The detected objects.
"""
detect_partial = partial(self._detector,
max_detections=max_detections,
orientations=orientations,
scale_factor=scale_factor,
stride_factor=stride_factor,
min_size=min_size,
confidence_cutoff=confidence_cutoff)
return detect(detect_partial,
image,
greyscale=True,
image_diagonal=image_diagonal,
group_prefix=group_prefix)
def test_passing_uint8_image_greyscale_raises():
takeo_copy = takeo_uint8.copy()
pcs = detect(fake_detector, takeo_copy, greyscale=True)
def __call__(self, image, image_diagonal=None, group_prefix='pico',
max_detections=100, orientations=0.0, degrees=True,
scale_factor=1.2, stride_factor=0.1, min_size=100,
confidence_cutoff=3.0, axis_aligned_bb=True):
r"""
Perform a detection using the cached pico detector.
The detections will also be attached to the image as landmarks.
Parameters
----------
image : `menpo.image.Image`
A Menpo image to detect. The bounding boxes of the detected objects
will be attached to this image.
image_diagonal : `int`, optional
The total size of the diagonal of the image that should be used for
detection. This is useful for scaling images up and down for
detection.
group_prefix : `str`, optional
The prefix string to be appended to each each landmark group that is
stored on the image. Each detection will be stored as group_prefix_#
where # is a count starting from 0.
max_detections : `int`, optional
The maximum number of detections to return.
orientations : list of `float`s or `float`, optional
The orientations of the cascades to use. ``0.0`` will perform an
axis aligned detection. Values greater than ``0.0`` will perform
detections of the cascade rotated counterclockwise around a unit
circle.
If a list is passed, each item should be an orientation in
either radians or degrees around the unit circle, with ``0.0``
being axis aligned.
degrees : `bool`, optional
If ``True``, the ``orientations`` parameter is treated as
rotations counterclockwise in degrees rather than radians.
scale_factor : `float`, optional
The ratio to increase the cascade window at every iteration. Must
be greater than 1.0
stride_factor : `float`, optional
The ratio to decrease the window step by at every iteration. Must be
less than 1.0, optional
min_size : `float`, optional
The minimum size in pixels (diameter of the detection circle) that a
face can be. This is the starting cascade window size.
confidence_cutoff : `float`, optional
The confidence value to trim the detections with. Any detections
with confidence less than the cutoff will be discarded.
axis_aligned_bb : `bool`, optional
If ``True``, the returned detections will be axis aligned,
regardless of which orientation they were detected at.
If ``False``, the returned bounding box will be rotated by the
orientation detected.
Returns
------
bounding_boxes : `list` of `menpo.shape.PointDirectedGraph`
The detected objects.
"""
if degrees:
# Cypico expects Radians
orientations = np.deg2rad(orientations)
detect_partial = partial(self._detector, max_detections=max_detections,
orientations=orientations,
scale_factor=scale_factor,
stride_factor=stride_factor, min_size=min_size,
confidence_cutoff=confidence_cutoff,
axis_aligned_bb=axis_aligned_bb)
return detect(detect_partial, image, greyscale=True,
image_diagonal=image_diagonal, group_prefix=group_prefix)
def test_passing_uint8_image():
takeo_copy = takeo_uint8.copy()
pcs = detect(fake_detector, takeo_copy, greyscale=False)
assert len(pcs) == 1
assert takeo_copy.n_channels == 3
assert takeo_copy.landmarks['object_0'].n_points == 4
def test_passing_uint8_image_greyscale():
takeo_copy = takeo_uint8.copy()
pcs = detect(fake_detector, takeo_copy, greyscale=True)
assert len(pcs) == 1
def test_passing_uint8_image_greyscale():
takeo_copy = takeo_uint8.copy()
pcs = detect(fake_detector, takeo_copy, greyscale=True)
assert len(pcs) == 1
def test_passing_uint8_greyscale_image_greyscale_pass_through():
fake_image = MagicMock()
fake_image.n_channels = 1
pcs = detect(fake_detector, fake_image, greyscale=True)
assert len(pcs) == 1
fake_image.as_greyscale.assert_not_called()
def test_passing_uint8_rgb_image_greyscale_no_convert():
fake_image = MagicMock()
fake_image.n_channels = 3
pcs = detect(fake_detector, fake_image, greyscale=True)
assert len(pcs) == 1
fake_image.as_greyscale.assert_called_once_with(mode='luminosity')
def test_passing_uint8_greyscale_image_greyscale_pass_through():
fake_image = MagicMock()
fake_image.n_channels = 1
pcs = detect(fake_detector, fake_image, greyscale=True)
assert len(pcs) == 1
fake_image.as_greyscale.assert_not_called()
def test_passing_uint8_image_greyscale_raises():
takeo_copy = takeo_uint8.copy()
pcs = detect(fake_detector, takeo_copy, greyscale=True)