help='RGB image directory')
args = parser.parse_args()
print(args)
strideNet = 8
Transform = outil.Homography
nbPoint = 4
torch.manual_seed(1000)
np.random.seed(1000)
## Loading model
# Define Networks
network = {
'netFeatCoarse': model.FeatureExtractor(),
'netCorr': model.CorrNeigh(args.kernelSize),
'netFlowCoarse': model.NetFlowCoarse(args.kernelSize),
'netMatch': model.NetMatchability(args.kernelSize),
}
for key in list(network.keys()):
network[key].cuda()
typeData = torch.cuda.FloatTensor
# loading Network
if args.resumePth:
param = torch.load(args.resumePth)
msg = 'Loading pretrained model from {}'.format(args.resumePth)
print(msg)
for key in list(param.keys()):
import cv2
from matplotlib import pyplot as plt
import trimesh
from getResults import opencv_decompose, _getFlow, matches_from_flow
from run_point_cloud import compute_and_save
sys.path.append('../')
Transform = outil.Homography
nbPoint = 4
## Loading model
# Define Networks
network = {
'netFeatCoarse': model.FeatureExtractor(),
'netCorr': model.CorrNeigh(kernelSize),
'netFlowCoarse': model.NetFlowCoarse(kernelSize),
'netMatch': model.NetMatchability(kernelSize),
}
if use_cuda:
device = torch.device("cuda")
for key in list(network.keys()):
network[key].cuda()
else:
device = torch.device("cpu")
# loading Network
param = torch.load(resumePth)
msg = 'Loading pretrained model from {}'.format(resumePth)