talksik/mediapipe
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Project import generated by Copybara.

Browse Source 
GitOrigin-RevId: 373e3ac1e5839befd95bf7d73ceff3c5f1171969
This commit is contained in:
MediaPipe Team
2021-10-06 14:27:49 -07:00
committed by jqtang
parent 137e1cc763
commit 33d683c671
153 changed files with 7871 additions and 1349 deletions
Download Patch File Download Diff File Expand all files Collapse all files
mediapipe/python/solutions/drawing_styles.py
+17
View File
@@ -183,6 +183,23 @@ def get_default_face_mesh_tesselation_style() -> DrawingSpec:
return DrawingSpec(color=_GRAY, thickness=_THICKNESS_TESSELATION)
def get_default_face_mesh_iris_connections_style(
) -> Mapping[Tuple[int, int], DrawingSpec]:
"""Returns the default face mesh iris connections drawing style.
Returns:
A mapping from each iris connection to its default drawing spec.
"""
face_mesh_iris_connections_style = {}
left_spec = DrawingSpec(color=_GREEN, thickness=_THICKNESS_CONTOURS)
for connection in face_mesh_connections.FACEMESH_LEFT_IRIS:
face_mesh_iris_connections_style[connection] = left_spec
right_spec = DrawingSpec(color=_RED, thickness=_THICKNESS_CONTOURS)
for connection in face_mesh_connections.FACEMESH_RIGHT_IRIS:
face_mesh_iris_connections_style[connection] = right_spec
return face_mesh_iris_connections_style
def get_default_pose_landmarks_style() -> Mapping[int, DrawingSpec]:
"""Returns the default pose landmarks drawing style.
mediapipe/python/solutions/drawing_utils.py
+3
View File
@@ -138,9 +138,12 @@ def draw_landmarks(
landmark_drawing_spec: Either a DrawingSpec object or a mapping from
hand landmarks to the DrawingSpecs that specifies the landmarks' drawing
settings such as color, line thickness, and circle radius.
If this argument is explicitly set to None, no landmarks will be drawn.
connection_drawing_spec: Either a DrawingSpec object or a mapping from
hand connections to the DrawingSpecs that specifies the
connections' drawing settings such as color and line thickness.
If this argument is explicitly set to None, no landmark connections will
be drawn.
Raises:
ValueError: If one of the followings:
mediapipe/python/solutions/face_detection.py
+3 -3
View File
@@ -28,8 +28,8 @@ from mediapipe.calculators.util import non_max_suppression_calculator_pb2
# pylint: enable=unused-import
from mediapipe.python.solution_base import SolutionBase
SHORT_RANGE_GRAPH_FILE_PATH = 'mediapipe/modules/face_detection/face_detection_short_range_cpu.binarypb'
FULL_RANGE_GRAPH_FILE_PATH = 'mediapipe/modules/face_detection/face_detection_full_range_cpu.binarypb'
_SHORT_RANGE_GRAPH_FILE_PATH = 'mediapipe/modules/face_detection/face_detection_short_range_cpu.binarypb'
_FULL_RANGE_GRAPH_FILE_PATH = 'mediapipe/modules/face_detection/face_detection_full_range_cpu.binarypb'
def get_key_point(
@@ -83,7 +83,7 @@ class FaceDetection(SolutionBase):
https://solutions.mediapipe.dev/face_detection#model_selection.
"""
binary_graph_path = FULL_RANGE_GRAPH_FILE_PATH if model_selection == 1 else SHORT_RANGE_GRAPH_FILE_PATH
binary_graph_path = _FULL_RANGE_GRAPH_FILE_PATH if model_selection == 1 else _SHORT_RANGE_GRAPH_FILE_PATH
subgraph_name = 'facedetectionfullrangecommon' if model_selection == 1 else 'facedetectionshortrangecommon'
super().__init__(
mediapipe/python/solutions/face_mesh.py
+20 -13
View File
@@ -12,14 +12,14 @@
# See the License for the specific language governing permissions and
# limitations under the License.
"""MediaPipe FaceMesh."""
"""MediaPipe Face Mesh."""
from typing import NamedTuple
import numpy as np
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
# pylint: disable=unused-import
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
from mediapipe.calculators.core import gate_calculator_pb2
from mediapipe.calculators.core import split_vector_calculator_pb2
from mediapipe.calculators.tensor import image_to_tensor_calculator_pb2
@@ -30,6 +30,7 @@ from mediapipe.calculators.tensor import tensors_to_landmarks_calculator_pb2
from mediapipe.calculators.tflite import ssd_anchors_calculator_pb2
from mediapipe.calculators.util import association_calculator_pb2
from mediapipe.calculators.util import detections_to_rects_calculator_pb2
from mediapipe.calculators.util import landmarks_refinement_calculator_pb2
from mediapipe.calculators.util import logic_calculator_pb2
from mediapipe.calculators.util import non_max_suppression_calculator_pb2
from mediapipe.calculators.util import rect_transformation_calculator_pb2
@@ -39,22 +40,26 @@ from mediapipe.python.solution_base import SolutionBase
# pylint: disable=unused-import
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_CONTOURS
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_FACE_OVAL
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_IRISES
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_LEFT_EYE
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_LEFT_EYEBROW
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_LEFT_IRIS
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_LIPS
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_RIGHT_EYE
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_RIGHT_EYEBROW
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_RIGHT_IRIS
from mediapipe.python.solutions.face_mesh_connections import FACEMESH_TESSELATION
# pylint: enable=unused-import
BINARYPB_FILE_PATH = 'mediapipe/modules/face_landmark/face_landmark_front_cpu.binarypb'
FACEMESH_NUM_LANDMARKS = 468
FACEMESH_NUM_LANDMARKS_WITH_IRISES = 478
_BINARYPB_FILE_PATH = 'mediapipe/modules/face_landmark/face_landmark_front_cpu.binarypb'
class FaceMesh(SolutionBase):
"""MediaPipe FaceMesh.
"""MediaPipe Face Mesh.
MediaPipe FaceMesh processes an RGB image and returns the face landmarks on
MediaPipe Face Mesh processes an RGB image and returns the face landmarks on
each detected face.
Please refer to https://solutions.mediapipe.dev/face_mesh#python-solution-api
@@ -64,9 +69,10 @@ class FaceMesh(SolutionBase):
def __init__(self,
static_image_mode=False,
max_num_faces=1,
refine_landmarks=False,
min_detection_confidence=0.5,
min_tracking_confidence=0.5):
"""Initializes a MediaPipe FaceMesh object.
"""Initializes a MediaPipe Face Mesh object.
Args:
static_image_mode: Whether to treat the input images as a batch of static
@@ -74,6 +80,10 @@ class FaceMesh(SolutionBase):
https://solutions.mediapipe.dev/face_mesh#static_image_mode.
max_num_faces: Maximum number of faces to detect. See details in
https://solutions.mediapipe.dev/face_mesh#max_num_faces.
refine_landmarks: Whether to further refine the landmark coordinates
around the eyes and lips, and output additional landmarks around the
irises. Default to False. See details in
https://solutions.mediapipe.dev/face_mesh#refine_landmarks.
min_detection_confidence: Minimum confidence value ([0.0, 1.0]) for face
detection to be considered successful. See details in
https://solutions.mediapipe.dev/face_mesh#min_detection_confidence.
@@ -82,16 +92,13 @@ class FaceMesh(SolutionBase):
https://solutions.mediapipe.dev/face_mesh#min_tracking_confidence.
"""
super().__init__(
binary_graph_path=BINARYPB_FILE_PATH,
binary_graph_path=_BINARYPB_FILE_PATH,
side_inputs={
'num_faces': max_num_faces,
'with_attention': refine_landmarks,
'use_prev_landmarks': not static_image_mode,
},
calculator_params={
'ConstantSidePacketCalculator.packet': [
constant_side_packet_calculator_pb2
.ConstantSidePacketCalculatorOptions.ConstantSidePacket(
bool_value=not static_image_mode)
],
'facedetectionshortrangecpu__facedetectionshortrangecommon__TensorsToDetectionsCalculator.min_score_thresh':
min_detection_confidence,
'facelandmarkcpu__ThresholdingCalculator.threshold':
mediapipe/python/solutions/face_mesh_connections.py
+8
View File
@@ -29,6 +29,9 @@ FACEMESH_LEFT_EYE = frozenset([(263, 249), (249, 390), (390, 373), (373, 374),
(263, 466), (466, 388), (388, 387), (387, 386),
(386, 385), (385, 384), (384, 398), (398, 362)])
FACEMESH_LEFT_IRIS = frozenset([(474, 475), (475, 476), (476, 477),
(477, 474)])
FACEMESH_LEFT_EYEBROW = frozenset([(276, 283), (283, 282), (282, 295),
(295, 285), (300, 293), (293, 334),
(334, 296), (296, 336)])
@@ -41,6 +44,9 @@ FACEMESH_RIGHT_EYE = frozenset([(33, 7), (7, 163), (163, 144), (144, 145),
FACEMESH_RIGHT_EYEBROW = frozenset([(46, 53), (53, 52), (52, 65), (65, 55),
(70, 63), (63, 105), (105, 66), (66, 107)])
FACEMESH_RIGHT_IRIS = frozenset([(469, 470), (470, 471), (471, 472),
(472, 469)])
FACEMESH_FACE_OVAL = frozenset([(10, 338), (338, 297), (297, 332), (332, 284),
(284, 251), (251, 389), (389, 356), (356, 454),
(454, 323), (323, 361), (361, 288), (288, 397),
@@ -56,6 +62,8 @@ FACEMESH_CONTOURS = frozenset().union(*[
FACEMESH_RIGHT_EYEBROW, FACEMESH_FACE_OVAL
])
FACEMESH_IRISES = frozenset().union(*[FACEMESH_LEFT_IRIS, FACEMESH_RIGHT_IRIS])
FACEMESH_TESSELATION = frozenset([
(127, 34), (34, 139), (139, 127), (11, 0), (0, 37), (37, 11),
(232, 231), (231, 120), (120, 232), (72, 37), (37, 39), (39, 72),
mediapipe/python/solutions/face_mesh_test.py
+41 -6
View File
@@ -67,10 +67,24 @@ EYE_INDICES_TO_LANDMARKS = {
398: [432, 175]
}
IRIS_INDICES_TO_LANDMARKS = {
468: [362, 175],
469: [371, 175],
470: [362, 167],
471: [354, 175],
472: [363, 182],
473: [449, 174],
474: [458, 174],
475: [449, 167],
476: [440, 174],
477: [449, 181]
}
class FaceMeshTest(parameterized.TestCase):
def _annotate(self, frame: np.ndarray, results: NamedTuple, idx: int):
def _annotate(self, frame: np.ndarray, results: NamedTuple, idx: int,
draw_iris: bool):
for face_landmarks in results.multi_face_landmarks:
mp_drawing.draw_landmarks(
frame,
@@ -86,6 +100,14 @@ class FaceMeshTest(parameterized.TestCase):
landmark_drawing_spec=None,
connection_drawing_spec=drawing_styles
.get_default_face_mesh_contours_style())
if draw_iris:
mp_drawing.draw_landmarks(
frame,
face_landmarks,
mp_faces.FACEMESH_IRISES,
landmark_drawing_spec=None,
connection_drawing_spec=drawing_styles
.get_default_face_mesh_iris_connections_style())
path = os.path.join(tempfile.gettempdir(), self.id().split('.')[-1] +
'_frame_{}.png'.format(idx))
cv2.imwrite(path, frame)
@@ -103,22 +125,29 @@ class FaceMeshTest(parameterized.TestCase):
results = faces.process(image)
self.assertIsNone(results.multi_face_landmarks)
@parameterized.named_parameters(('static_image_mode', True, 1),
('video_mode', False, 5))
def test_face(self, static_image_mode: bool, num_frames: int):
@parameterized.named_parameters(
('static_image_mode_no_attention', True, False, 5),
('static_image_mode_with_attention', True, True, 5),
('streaming_mode_no_attention', False, False, 10),
('streaming_mode_with_attention', False, True, 10))
def test_face(self, static_image_mode: bool, refine_landmarks: bool,
num_frames: int):
image_path = os.path.join(os.path.dirname(__file__),
'testdata/portrait.jpg')
image = cv2.imread(image_path)
rows, cols, _ = image.shape
with mp_faces.FaceMesh(
static_image_mode=static_image_mode,
refine_landmarks=refine_landmarks,
min_detection_confidence=0.5) as faces:
for idx in range(num_frames):
results = faces.process(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
self._annotate(image.copy(), results, idx)
self._annotate(image.copy(), results, idx, refine_landmarks)
multi_face_landmarks = []
for landmarks in results.multi_face_landmarks:
self.assertLen(landmarks.landmark, 468)
self.assertLen(
landmarks.landmark, mp_faces.FACEMESH_NUM_LANDMARKS_WITH_IRISES
if refine_landmarks else mp_faces.FACEMESH_NUM_LANDMARKS)
x = [landmark.x * cols for landmark in landmarks.landmark]
y = [landmark.y * rows for landmark in landmarks.landmark]
face_landmarks = np.column_stack((x, y))
@@ -129,6 +158,12 @@ class FaceMeshTest(parameterized.TestCase):
prediction_error = np.abs(
np.asarray(multi_face_landmarks[0][eye_idx]) - np.asarray(gt_lds))
npt.assert_array_less(prediction_error, DIFF_THRESHOLD)
if refine_landmarks:
for iris_idx, gt_lds in IRIS_INDICES_TO_LANDMARKS.items():
prediction_error = np.abs(
np.asarray(multi_face_landmarks[0][iris_idx]) -
np.asarray(gt_lds))
npt.assert_array_less(prediction_error, DIFF_THRESHOLD)
if __name__ == '__main__':
mediapipe/python/solutions/hands.py
+4 -8
View File
@@ -19,8 +19,8 @@ from typing import NamedTuple
import numpy as np
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
# pylint: disable=unused-import
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
from mediapipe.calculators.core import gate_calculator_pb2
from mediapipe.calculators.core import split_vector_calculator_pb2
from mediapipe.calculators.tensor import image_to_tensor_calculator_pb2
@@ -67,7 +67,7 @@ class HandLandmark(enum.IntEnum):
PINKY_TIP = 20
BINARYPB_FILE_PATH = 'mediapipe/modules/hand_landmark/hand_landmark_tracking_cpu.binarypb'
_BINARYPB_FILE_PATH = 'mediapipe/modules/hand_landmark/hand_landmark_tracking_cpu.binarypb'
class Hands(SolutionBase):
@@ -107,16 +107,12 @@ class Hands(SolutionBase):
https://solutions.mediapipe.dev/hands#min_tracking_confidence.
"""
super().__init__(
binary_graph_path=BINARYPB_FILE_PATH,
binary_graph_path=_BINARYPB_FILE_PATH,
side_inputs={
'num_hands': max_num_hands,
'use_prev_landmarks': not static_image_mode,
},
calculator_params={
'ConstantSidePacketCalculator.packet': [
constant_side_packet_calculator_pb2
.ConstantSidePacketCalculatorOptions.ConstantSidePacket(
bool_value=not static_image_mode)
],
'palmdetectioncpu__TensorsToDetectionsCalculator.min_score_thresh':
min_detection_confidence,
'handlandmarkcpu__ThresholdingCalculator.threshold':
mediapipe/python/solutions/hands_test.py
+6 -6
View File
@@ -32,20 +32,20 @@ from mediapipe.python.solutions import hands as mp_hands
TEST_IMAGE_PATH = 'mediapipe/python/solutions/testdata'
DIFF_THRESHOLD = 15 # pixels
EXPECTED_HAND_COORDINATES_PREDICTION = [[[144, 345], [211, 323], [257, 286],
DIFF_THRESHOLD = 20 # pixels
EXPECTED_HAND_COORDINATES_PREDICTION = [[[138, 343], [211, 330], [257, 286],
[289, 237], [322, 203], [219, 216],
[238, 138], [249, 90], [253, 51],
[177, 204], [184, 115], [187, 60],
[185, 19], [138, 208], [131, 127],
[124, 77], [117, 36], [106, 222],
[92, 159], [79, 124], [68, 93]],
[[577, 37], [504, 56], [459, 94],
[429, 146], [397, 182], [496, 167],
[[580, 36], [504, 50], [459, 94],
[429, 146], [397, 182], [507, 167],
[479, 245], [469, 292], [464, 330],
[540, 177], [534, 265], [533, 319],
[545, 180], [534, 265], [533, 319],
[536, 360], [581, 172], [587, 252],
[593, 304], [599, 346], [615, 157],
[593, 304], [599, 346], [615, 168],
[628, 223], [638, 258], [648, 288]]]
mediapipe/python/solutions/holistic.py
+19 -12
View File
@@ -17,10 +17,10 @@ from typing import NamedTuple
import numpy as np
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
# The following imports are needed because python pb2 silently discards
# unknown protobuf fields.
# pylint: disable=unused-import
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
from mediapipe.calculators.core import gate_calculator_pb2
from mediapipe.calculators.core import split_vector_calculator_pb2
from mediapipe.calculators.tensor import image_to_tensor_calculator_pb2
@@ -49,7 +49,7 @@ from mediapipe.python.solutions.pose import PoseLandmark
from mediapipe.python.solutions.pose_connections import POSE_CONNECTIONS
# pylint: enable=unused-import
BINARYPB_FILE_PATH = 'mediapipe/modules/holistic_landmark/holistic_landmark_cpu.binarypb'
_BINARYPB_FILE_PATH = 'mediapipe/modules/holistic_landmark/holistic_landmark_cpu.binarypb'
def _download_oss_pose_landmark_model(model_complexity):
@@ -78,6 +78,8 @@ class Holistic(SolutionBase):
static_image_mode=False,
model_complexity=1,
smooth_landmarks=True,
enable_segmentation=False,
smooth_segmentation=True,
min_detection_confidence=0.5,
min_tracking_confidence=0.5):
"""Initializes a MediaPipe Holistic object.
@@ -91,6 +93,11 @@ class Holistic(SolutionBase):
smooth_landmarks: Whether to filter landmarks across different input
images to reduce jitter. See details in
https://solutions.mediapipe.dev/holistic#smooth_landmarks.
enable_segmentation: Whether to predict segmentation mask. See details in
https://solutions.mediapipe.dev/holistic#enable_segmentation.
smooth_segmentation: Whether to filter segmentation across different input
images to reduce jitter. See details in
https://solutions.mediapipe.dev/holistic#smooth_segmentation.
min_detection_confidence: Minimum confidence value ([0.0, 1.0]) for person
detection to be considered successful. See details in
https://solutions.mediapipe.dev/holistic#min_detection_confidence.
@@ -100,18 +107,16 @@ class Holistic(SolutionBase):
"""
_download_oss_pose_landmark_model(model_complexity)
super().__init__(
binary_graph_path=BINARYPB_FILE_PATH,
binary_graph_path=_BINARYPB_FILE_PATH,
side_inputs={
'model_complexity': model_complexity,
'smooth_landmarks': smooth_landmarks and not static_image_mode,
'smooth_segmentation': not static_image_mode,
'enable_segmentation': enable_segmentation,
'smooth_segmentation':
smooth_segmentation and not static_image_mode,
'use_prev_landmarks': not static_image_mode,
},
calculator_params={
'poselandmarkcpu__ConstantSidePacketCalculator.packet': [
constant_side_packet_calculator_pb2
.ConstantSidePacketCalculatorOptions.ConstantSidePacket(
bool_value=not static_image_mode)
],
'poselandmarkcpu__posedetectioncpu__TensorsToDetectionsCalculator.min_score_thresh':
min_detection_confidence,
'poselandmarkcpu__poselandmarkbyroicpu__tensorstoposelandmarksandsegmentation__ThresholdingCalculator.threshold':
@@ -119,7 +124,7 @@ class Holistic(SolutionBase):
},
outputs=[
'pose_landmarks', 'pose_world_landmarks', 'left_hand_landmarks',
'right_hand_landmarks', 'face_landmarks'
'right_hand_landmarks', 'face_landmarks', 'segmentation_mask'
])
def process(self, image: np.ndarray) -> NamedTuple:
@@ -133,8 +138,8 @@ class Holistic(SolutionBase):
ValueError: If the input image is not three channel RGB.
Returns:
A NamedTuple that has five fields describing the landmarks on the most
prominate person detected:
A NamedTuple with fields describing the landmarks on the most prominate
person detected:
1) "pose_landmarks" field that contains the pose landmarks.
2) "pose_world_landmarks" field that contains the pose landmarks in
real-world 3D coordinates that are in meters with the origin at the
@@ -142,6 +147,8 @@ class Holistic(SolutionBase):
3) "left_hand_landmarks" field that contains the left-hand landmarks.
4) "right_hand_landmarks" field that contains the right-hand landmarks.
5) "face_landmarks" field that contains the face landmarks.
6) "segmentation_mask" field that contains the segmentation mask if
"enable_segmentation" is set to true.
"""
results = super().process(input_data={'image': image})
mediapipe/python/solutions/objectron.py
+4 -8
View File
@@ -20,8 +20,8 @@ from typing import List, Tuple, NamedTuple, Optional
import attr
import numpy as np
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
# pylint: disable=unused-import
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
from mediapipe.calculators.core import gate_calculator_pb2
from mediapipe.calculators.core import split_vector_calculator_pb2
from mediapipe.calculators.tensor import image_to_tensor_calculator_pb2
@@ -75,7 +75,7 @@ class BoxLandmark(enum.IntEnum):
BACK_TOP_RIGHT = 7
FRONT_TOP_RIGHT = 8
BINARYPB_FILE_PATH = 'mediapipe/modules/objectron/objectron_cpu.binarypb'
_BINARYPB_FILE_PATH = 'mediapipe/modules/objectron/objectron_cpu.binarypb'
BOX_CONNECTIONS = frozenset([
(BoxLandmark.BACK_BOTTOM_LEFT, BoxLandmark.FRONT_BOTTOM_LEFT),
(BoxLandmark.BACK_BOTTOM_LEFT, BoxLandmark.BACK_TOP_LEFT),
@@ -216,18 +216,14 @@ class Objectron(SolutionBase):
# Create and init model.
model = get_model_by_name(model_name)
super().__init__(
binary_graph_path=BINARYPB_FILE_PATH,
binary_graph_path=_BINARYPB_FILE_PATH,
side_inputs={
'box_landmark_model_path': model.model_path,
'allowed_labels': model.label_name,
'max_num_objects': max_num_objects,
'use_prev_landmarks': not static_image_mode,
},
calculator_params={
'ConstantSidePacketCalculator.packet': [
constant_side_packet_calculator_pb2
.ConstantSidePacketCalculatorOptions.ConstantSidePacket(
bool_value=not static_image_mode)
],
('objectdetectionoidv4subgraph'
'__TensorsToDetectionsCalculator.min_score_thresh'):
min_detection_confidence,
mediapipe/python/solutions/pose.py
+8 -10
View File
@@ -19,10 +19,10 @@ from typing import NamedTuple
import numpy as np
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
# The following imports are needed because python pb2 silently discards
# unknown protobuf fields.
# pylint: disable=unused-import
from mediapipe.calculators.core import constant_side_packet_calculator_pb2
from mediapipe.calculators.core import gate_calculator_pb2
from mediapipe.calculators.core import split_vector_calculator_pb2
from mediapipe.calculators.image import warp_affine_calculator_pb2
@@ -87,7 +87,7 @@ class PoseLandmark(enum.IntEnum):
RIGHT_FOOT_INDEX = 32
BINARYPB_FILE_PATH = 'mediapipe/modules/pose_landmark/pose_landmark_cpu.binarypb'
_BINARYPB_FILE_PATH = 'mediapipe/modules/pose_landmark/pose_landmark_cpu.binarypb'
def _download_oss_pose_landmark_model(model_complexity):
@@ -144,20 +144,16 @@ class Pose(SolutionBase):
"""
_download_oss_pose_landmark_model(model_complexity)
super().__init__(
binary_graph_path=BINARYPB_FILE_PATH,
binary_graph_path=_BINARYPB_FILE_PATH,
side_inputs={
'model_complexity': model_complexity,
'smooth_landmarks': smooth_landmarks and not static_image_mode,
'enable_segmentation': enable_segmentation,
'smooth_segmentation':
smooth_segmentation and not static_image_mode,
'use_prev_landmarks': not static_image_mode,
},
calculator_params={
'ConstantSidePacketCalculator.packet': [
constant_side_packet_calculator_pb2
.ConstantSidePacketCalculatorOptions.ConstantSidePacket(
bool_value=not static_image_mode)
],
'posedetectioncpu__TensorsToDetectionsCalculator.min_score_thresh':
min_detection_confidence,
'poselandmarkbyroicpu__tensorstoposelandmarksandsegmentation__ThresholdingCalculator.threshold':
@@ -176,12 +172,14 @@ class Pose(SolutionBase):
ValueError: If the input image is not three channel RGB.
Returns:
A NamedTuple that has two fields describing the landmarks on the most
prominate person detected:
A NamedTuple with fields describing the landmarks on the most prominate
person detected:
1) "pose_landmarks" field that contains the pose landmarks.
2) "pose_world_landmarks" field that contains the pose landmarks in
real-world 3D coordinates that are in meters with the origin at the
center between hips.
3) "segmentation_mask" field that contains the segmentation mask if
"enable_segmentation" is set to true.
"""
results = super().process(input_data={'image': image})
mediapipe/python/solutions/selfie_segmentation.py
+2 -2
View File
@@ -29,7 +29,7 @@ from mediapipe.framework.tool import switch_container_pb2
from mediapipe.python.solution_base import SolutionBase
BINARYPB_FILE_PATH = 'mediapipe/modules/selfie_segmentation/selfie_segmentation_cpu.binarypb'
_BINARYPB_FILE_PATH = 'mediapipe/modules/selfie_segmentation/selfie_segmentation_cpu.binarypb'
class SelfieSegmentation(SolutionBase):
@@ -52,7 +52,7 @@ class SelfieSegmentation(SolutionBase):
https://solutions.mediapipe.dev/selfie_segmentation#model_selection.
"""
super().__init__(
binary_graph_path=BINARYPB_FILE_PATH,
binary_graph_path=_BINARYPB_FILE_PATH,
side_inputs={
'model_selection': model_selection,
},