def SIFTgrain(self, octave):
path_analyse = '/CT_analysesClassification/' + octave
octaves = 6
initial_sigma = 1.1
gauss = GaussianSmoothing3D(self.path + path_analyse, octaves).smoothing(initial_sigma)
dimension = 3
dog = DoG(self.path + path_analyse, dimension)
dog.apply()
pathDoG = '/3DDoG/'
local_extrema = LocalExterma3D(self.path + path_analyse + pathDoG, self.path + path_analyse, True)
local_extrema.find()
path_local_extrema = '/3DLocalExtremum/'
extrema = ExtremaSpace3D(self.path + path_analyse + pathDoG + path_local_extrema)
extrema.find()
Hessian = HessianMatrix(5.0)
Hessian.HessianElimination(self.path + path_analyse)
keypointorientation = KeyPointOrientation(self.path + path_analyse)
keypointorientation.apply()
path_keypoint_orientation = '/KeyPointsOrientation/'
images = ReadImage(self.path + path_analyse + path_keypoint_orientation).openImage()
for im in images:
rotateImage(im, 10, self.path + path_analyse).apply()
path_discriptor = '/Descriptor3D/'
keydis = KeypointsFeatures(self.path + path_analyse + path_discriptor, self.path + path_analyse).apply()
def SIFTgrain(self, octave):
path_analyse = '/CT_analysesClassification/' + octave
octaves = 6
initial_sigma = 1.1
gauss = GaussianSmoothing3D(self.path + path_analyse,
octaves).smoothing(initial_sigma)
dimension = 3
dog = DoG(self.path + path_analyse, dimension)
dog.apply()
pathDoG = '/3DDoG/'
local_extrema = LocalExterma3D(self.path + path_analyse + pathDoG,
self.path + path_analyse, True)
local_extrema.find()
path_local_extrema = '/3DLocalExtremum/'
extrema = ExtremaSpace3D(self.path + path_analyse + pathDoG +
path_local_extrema)
extrema.find()
Hessian = HessianMatrix(5.0)
Hessian.HessianElimination(self.path + path_analyse)
keypointorientation = KeyPointOrientation(self.path + path_analyse)
keypointorientation.apply()
path_keypoint_orientation = '/KeyPointsOrientation/'
images = ReadImage(self.path + path_analyse +
path_keypoint_orientation).openImage()
for im in images:
rotateImage(im, 10, self.path + path_analyse).apply()
path_discriptor = '/Descriptor3D/'
keydis = KeypointsFeatures(self.path + path_analyse + path_discriptor,
self.path + path_analyse).apply()
class DoG2DTest(unittest.TestCase):
def setUp(self):
self.path = './test_data/1_nd/CT_analyses/'
self.dog = DoG(self.path, 2)
def test_smoothing(self):
self.dog.apply()
def test_min(self):
path = './test_data/1_nd/CT_analyses/2DDoG/'
self.ReadImage = ReadImage(path)
list_of_image = self.ReadImage.openImage()
for z in list_of_image:
print min(z.Image3D),max(z.Image3D),z.sigma
def test_show(self):
path = './test_data/1_nd/CT_analyses/2DDoG/'
self.ReadImage = ReadImage(path)
list_of_image = self.ReadImage.openImage()
for Z in list_of_image:
visualization2D(Z.Image3D)
class DoG2DTest(unittest.TestCase):
def setUp(self):
self.path = './test_data/1_nd/CT_analyses/'
self.dog = DoG(self.path, 2)
def test_smoothing(self):
self.dog.apply()
def test_min(self):
path = './test_data/1_nd/CT_analyses/2DDoG/'
self.ReadImage = ReadImage(path)
list_of_image = self.ReadImage.openImage()
for z in list_of_image:
print min(z.Image3D), max(z.Image3D), z.sigma
def test_show(self):
path = './test_data/1_nd/CT_analyses/2DDoG/'
self.ReadImage = ReadImage(path)
list_of_image = self.ReadImage.openImage()
for Z in list_of_image:
visualization2D(Z.Image3D)
def setUp(self):
self.path = './test_data/1_nd/CT_analyses/'
self.dog = DoG(self.path, 3)
from DoG import DoG
import time
## Defining the communicating port
T_SIM=30;
DoG1=DoG(T_SIM='30',
input_ip=['localhost'],
input_port=['6666'],
output_port=['6665'],
mode_comm='RepReq')
start=time.time()
i=0
task_delay=list()
while time.time()-start<T_SIM:
#DoG1.REQ_array_image()
DoG1.REQ_array()
#a=round(DoG1.REQ_latency(),2)
#if i>4:
# average=round(sum(DoG1._delay_req[4:])/len(DoG1._delay_req[4:]),2)
# maxi=round(max(DoG1._delay_req[4:]),2)
# print("--- REQ instant --- : {0} ms --- REQ max --- : {1} ms , --- REQ avg --- {2} ms".format(a,maxi,average))
start=time.time()
DoG1.DoG1_save()
task_delay.append(round((time.time()-start)*1000,2))
DoG1.display_task_delay(task_delay)
DoG1.REP_array(DoG1.output_data)
layers.append(init_layers(total_time=total_time, C=20, H=54, W=54))
layer_4 = pool2().to('cuda')
layers.append(init_layers(total_time=total_time, C=20, H=10, W=10))
layer_5 = conv3().to('cuda')
layers.append(init_layers(total_time=total_time, C=20, H=6, W=6))
network = [layer_1, layer_2, layer_3, layer_4, layer_5]
network_len = len(network)
layer_filt_dimension = get_dim(network[0::2])
thresh = torch.Tensor([.5, .5, .5])
image_names = os.listdir(train_data)
data_len = len(image_names)
thds_per_dim = 10
for j in trange(data_len):
image = DoG(image_name=image_names[j])
image = np.reshape(image, (160000, ))
image = freq(15, image, 0.05)
image = torch.from_numpy(image).to('cuda')
image = image.float()
for i in range(stdp_params['max_iter']):
for t in range(total_time):
for l in range(network_len):
V = layers[l]['V'][t - 1, :, :, :]
S = layers[l]['S'][t, :, :, :]
K_inh = layers[l]['K_inh']
if network[l].__class__.__name__ == "conv1":
img = image[t, :, :, :].unsqueeze(0)
V = network[l](img)
V = V.permute(0, 3, 2, 1)
def setUp(self):
self.path = './test_data/1_nd/CT_analyses/'
self.dog = DoG(self.path, 3)