def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
# Create parameters class
if param is None:
self.setParam(PointRendParam())
else:
self.setParam(copy.deepcopy(param))
self.threshold = 0.5
self.MODEL_NAME_CONFIG = "pointrend_rcnn_R_50_FPN_3x_coco"
self.MODEL_NAME = "model_final_3c3198"
self.path_to_config = "/PointRend_git/configs/InstanceSegmentation/" + self.MODEL_NAME_CONFIG + ".yaml"
self.path_to_model = "/models/" + self.MODEL_NAME + ".pkl"
self.folder = os.path.dirname(os.path.realpath(__file__))
self.cfg = get_cfg()
add_pointrend_config(self.cfg)
self.cfg.merge_from_file(self.folder + self.path_to_config)
self.cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = self.threshold
self.cfg.MODEL.WEIGHTS = self.folder + self.path_to_model
self.loaded = False
self.deviceFrom = ""
# add output + set data type
self.setOutputDataType(core.IODataType.IMAGE_LABEL, 0)
self.addOutput(dataprocess.CImageIO(core.IODataType.IMAGE))
self.addOutput(dataprocess.CGraphicsOutput())
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
if param is None:
self.setParam(MaskRcnnParam())
else:
self.setParam(copy.deepcopy(param))
# get and set config model
self.threshold = 0.5
self.cfg = None
self.model_link = "COCO-InstanceSegmentation/mask_rcnn_X_101_32x8d_FPN_3x.yaml"
self.predictor = None
# add output + set data type
self.setOutputDataType(core.IODataType.IMAGE_LABEL, 0)
self.addOutput(dataprocess.CImageIO(core.IODataType.IMAGE))
self.addOutput(dataprocess.CGraphicsOutput())
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
self.model = None
self.names = None
self.colors = None
self.update = False
# Detect if we have a GPU available
self.device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# Add graphics output
self.addOutput(dataprocess.CGraphicsOutput())
# Add numeric output
self.addOutput(dataprocess.CNumericIO())
# Create parameters class
if param is None:
self.setParam(InferYoloV5Param())
else:
self.setParam(copy.deepcopy(param))
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
self.model = None
self.class_names = []
# Detect if we have a GPU available
self.device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# Remove graphics input
self.removeInput(1)
# Add graphics output
self.addOutput(dataprocess.CGraphicsOutput())
# Add numeric output
self.addOutput(dataprocess.CNumericIO())
# Create parameters class
if param is None:
self.setParam(MnasnetParam())
else:
self.setParam(copy.deepcopy(param))
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
if param is None:
self.setParam(TridentnetParam())
else:
self.setParam(copy.deepcopy(param))
# get and set config model
self.folder = os.path.dirname(os.path.realpath(__file__))
self.MODEL_NAME_CONFIG = "tridentnet_fast_R_101_C4_3x"
self.MODEL_NAME = "model_final_164568"
self.cfg = get_cfg()
add_tridentnet_config(self.cfg)
self.cfg.merge_from_file(self.folder + "/TridentNet_git/configs/" +
self.MODEL_NAME_CONFIG + ".yaml")
self.cfg.MODEL.WEIGHTS = self.folder + "/models/" + self.MODEL_NAME + ".pkl"
self.loaded = False
self.deviceFrom = ""
# add output
self.addOutput(dataprocess.CGraphicsOutput())
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
# Add graphics output
self.addOutput(dataprocess.CGraphicsOutput())
# Add numeric output
self.addOutput(dataprocess.CNumericIO())
# Network members
self.net = None
self.class_names = []
# Create parameters class
if param is None:
self.setParam(EmotionFerPlusParam())
else:
self.setParam(copy.deepcopy(param))
# Load class names
model_folder = os.path.dirname(os.path.realpath(__file__)) + "/models"
with open(model_folder + "/class_names") as f:
for row in f:
self.class_names.append(row[:-1])
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
self.model = None
self.class_names = []
# Detect if we have a GPU available
self.device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# Remove graphics input
self.removeInput(1)
# Segmentation mask output
self.setOutputDataType(core.IODataType.IMAGE_LABEL, 0)
# Result image
self.addOutput(dataprocess.CImageIO(core.IODataType.IMAGE))
# Add graphics output
self.addOutput(dataprocess.CGraphicsOutput())
# Add numeric output
self.addOutput(dataprocess.CNumericIO())
# Create parameters class
if param is None:
self.setParam(MaskRcnnParam())
else:
self.setParam(copy.deepcopy(param))
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
# Add input/output of the process here
self.setOutputDataType(core.IODataType.IMAGE_LABEL, 0)
self.addOutput(dataprocess.CImageIO(core.IODataType.IMAGE))
# Add graphics output
self.addOutput(dataprocess.CGraphicsOutput())
self.net = None
self.class_names = []
# Create parameters class
if param is None:
self.setParam(InferYolactParam())
else:
self.setParam(copy.deepcopy(param))
# Load class names
model_folder = os.path.dirname(os.path.realpath(__file__)) + "/models"
with open(model_folder + "/Coco_names.txt") as f:
for row in f:
self.class_names.append(row[:-1])
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
# Create parameters class
if param is None:
self.setParam(RetinanetParam())
else:
self.setParam(copy.deepcopy(param))
# get and set config model
self.LINK_MODEL = "COCO-Detection/retinanet_R_101_FPN_3x.yaml"
self.threshold = 0.5
self.cfg = get_cfg()
self.cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = self.threshold
self.cfg.merge_from_file(model_zoo.get_config_file(
self.LINK_MODEL)) # load config from file(.yaml)
self.cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url(
self.LINK_MODEL) # download the model (.pkl)
self.loaded = False
self.deviceFrom = ""
# add output
self.addOutput(dataprocess.CGraphicsOutput())
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
# Create parameters class
if param is None:
self.setParam(DensePoseParam())
else:
self.setParam(copy.deepcopy(param))
# get and set config model
self.folder = os.path.dirname(os.path.realpath(__file__))
self.MODEL_NAME_CONFIG = "densepose_rcnn_R_50_FPN_s1x"
self.MODEL_NAME = "model_final_162be9"
self.cfg = get_cfg()
add_densepose_config(self.cfg)
self.cfg.merge_from_file(self.folder + "/DensePose_git/configs/"+self.MODEL_NAME_CONFIG+".yaml") # load densepose_rcnn_R_101_FPN_d config from file(.yaml)
self.cfg.MODEL.WEIGHTS = self.folder + "/models/"+self.MODEL_NAME+".pkl" # load densepose_rcnn_R_101_FPN_d config from file(.pkl)
self.loaded = False
self.deviceFrom = ""
# add output graph
self.addOutput(dataprocess.CGraphicsOutput())
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
# Add graphics output
self.addOutput(dataprocess.CGraphicsOutput())
# Add numeric output
self.addOutput(dataprocess.CNumericIO())
# Create parameters class
if param is None:
self.setParam(CovidNetParam())
else:
self.setParam(copy.deepcopy(param))
param = self.getParam()
self.covid_model = CovidNet(model_path=param.model_path)
# Load class names
self.class_names = []
class_names_path = os.path.dirname(os.path.realpath(__file__)) + "/models/class_names"
with open(class_names_path) as f:
for row in f:
self.class_names.append(row[:-1])
def __init__(self, name, param):
dataprocess.C2dImageTask.__init__(self, name)
if param is None:
self.setParam(KeyptsRcnnParam())
else:
self.setParam(copy.deepcopy(param))
# get and set config model
self.LINK_MODEL = "COCO-Keypoints/keypoint_rcnn_X_101_32x8d_FPN_3x.yaml"
self.cfg = get_cfg()
self.threshold = 0.5
self.cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = self.threshold
self.cfg.merge_from_file(model_zoo.get_config_file(
self.LINK_MODEL)) # load config from file(.yaml)
self.cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url(
self.LINK_MODEL) # download the model (.pkl)
self.loaded = False
self.deviceFrom = ""
# add output
self.addOutput(dataprocess.CGraphicsOutput())
# keypoint threshold
self._KEYPOINT_THRESHOLD = 0.05