def run(self):
self.mask = Mask(self.frames[0])
if (PARAMS["verbose"]): self.mask.render("contour")
self.mask.computeDensity()
if (PARAMS["debug"]): self.mask.render("density")
self.currentcontour = self.mask.contour
RESULT = []
for i in range(0, len(self.frames)):
RESULT += [self.computeSimpleFlow(i)]
print(RESULT[0].shape, "SHAPE ", vid.getFPS(self.path), "FPS")
print([r.shape for r in RESULT])
shape = (int(RESULT[0].shape[0] * 2), int(RESULT[0].shape[1]))
print(shape)
out = cv2.VideoWriter('output.avi', cv2.VideoWriter_fourcc(*'XVID'),
vid.getFPS(self.path), (512, 512))
print(vid.getFPS(self.path), shape)
for f, res in zip(self.noblurframes, RESULT):
res3 = np.array([res, res, res]).transpose(1, 2, 0) * 255
newf = cv2.resize(np.uint8(np.concatenate([f, res3], axis=0)),
(512, 512))
#plt.imshow(newf[:, :, ::-1])
#plt.show()
out.write(newf)
out.release()
def print_new_dead_new_alive():
start_orbit = 21847
end_orbit = 21890
# start_orbit = 22820
# end_orbit = 22890
# start_orbit = 24023
# end_orbit = 24051
# start_orbit = 27229
# end_orbit = 27258
# start_orbit = 32726
# end_orbit = 32870
# start_orbit = 49230
# end_orbit = 49259
m1 = Mask()
m1.apply_pixel_window(394, 620)
m1.load_ascii(start_orbit)
m2 = Mask()
for orbit in range(start_orbit, end_orbit):
m2.load_ascii(orbit)
print(orbit, "new dead:", m1.get_new_dead(m2), "new alive:", m1.get_new_alive(m2))
return
def __init__(self, rect, mask):
self.rect = rect
if not self.rect.size == mask.get_size():
self.rect.size = mask.get_size()
Mask.__init__(self, mask.get_size())
self.from_array(mask.get_array())
def applyMedian(self):
mask = Mask('Mascaras/mask-median.txt')
mask.show('Median')
matrix = self.pbm.getMatrix()
data = matrix.map(moveMask(mask, matrix, median))
matrix.setMatrixData(data)
print('Applied Median')
def process(infile, group_name, day, timeid, copies=1):
mask = Mask((A4_WIDTH, A4_HEIGHT))
page = mask.apply_mask(infile)
create_title(page, group_name, day, timeid)
if not os.path.exists(os.path.join('png', day)):
os.mkdir(os.path.join('png', day))
png_file = os.path.join('png', day, '{}_{}.jpg'.format(timeid, tools.safe_filename(group_name)))
page.save(png_file, quality=75)
tools.print_image(png_file, copies)
def demoMaskCreation():
image = mycv2.loadImage(PARAMS["data"]["images"]["lena"])
image = mycv2.cvtGray(image)
mycv2.show(image)
mask = Mask(image)
mycv2.show(mask.mask)
mycv2.show(mask.hull)
print("Area", mask.area())
def process(infile, group_name, day, timeid, copies=1):
mask = Mask((A4_WIDTH, A4_HEIGHT))
page = mask.apply_mask(infile)
create_title(page, group_name, day, timeid)
tools.mkdir_p(os.path.join('png', day))
png_file = os.path.join(
'png', day, '{}_{}.jpg'.format(timeid,
tools.safe_filename(group_name)))
page.save(png_file, quality=75)
tools.print_image(png_file, copies)
def __init__(self, distance_threshold, depthmask_file, scenemask_file, scenemask_threshold, frame_threshold):
self.idxm = Identifier()
self.targets = {}
self.lost_targets = {}
#self.last_targets = set()
# thresholds
self.distance_threshold = distance_threshold
self.depthmask = cv2.imread(depthmask_file).mean(2) / 255.0
self.scenemask = Mask(scenemask_file, scenemask_threshold)
self.frame_threshold = frame_threshold
def compare_smoothmasks():
fname = input_mask_fname #"sdmf_pyxelmask.h5"
fid = h5py.File(fname, "r")
ds_orbits = fid["orbits"]
idx = np.argsort(ds_orbits[:])
orbits32 = ds_orbits[:][idx]
start_orbit = np.min(orbits32)
stop_orbit = np.max(orbits32)
orbit_range = stop_orbit - start_orbit + 1
#print(idx)
ds_combi = fid["combined"]
num_orbits = idx.size
a = np.empty((num_orbits,1024), dtype=np.float)
for i_orbit in range(num_orbits):
id_ = idx[i_orbit]
orbit = orbits32[i_orbit]
# a[orbit-start_orbit,:] = ds_combi[id_,:]
a[i_orbit,:] = ds_combi[id_,:]
fid.close()
print("3.2 read")
# print(a.shape)
# plt.cla()
# plt.imshow(a)
# plt.show()
# fname_out = "orbital_dbqm.h5"
# fid_out = h5py.File(fname_out, "w")
# ds = fid_out.create_dataset("data", a.shape, dtype=np.float)
# ds[:,:] = a
# fid_out.close()
# np.savetxt("orbital_dbqm.csv", a, delimiter=",")
m = Mask()
a_binary = np.ones((1400,1024), dtype=np.bool)
i_orbit = 0
orbits30 = np.empty(1400, dtype=np.int)
for orbit in range(42000,43400):
#print(orbit)
try:
m.load_sdmf30_crit(orbit, "combined", smooth=True)
except:
continue
orbits30[i_orbit] = orbit
a_binary[i_orbit,:] = m.mask
i_orbit += 1
print("3.0 read")
orbits30 = orbits30[0:i_orbit]
a_binary = a_binary[0:i_orbit,:]
for pixnr in range(1024):
plot(orbits32, a, orbits30, a_binary, pixnr)
return
def __init__( self ):
"""
Placeholder
"""
# increase system memory usage
os.environ['GDAL_CACHEMAX'] = '2048'
gdal.UseExceptions()
self._mask = Mask()
return
def get_program(filename):
program = []
mask = None
with open(filename, 'r') as f:
for line in f:
line = line[:-1]
if 'mask' in line:
if mask is not None:
program.append(mask)
mask = Mask(line)
elif 'mem' in line:
mask.add_mem(line)
program.append(mask)
return program
def process_pair_data(self, data_r1_path, data_r2_path, force_mask = None):
"""Used to read and process a pair of FASTQ data files.
Note that this parses the pair data into an in-memory SQLite
database, which on most modern systems will be fine except for the
largest input sets. If you hit memory issues, create a disk-based
SQLite DB via :class:`.db.PairDB` and then use
:meth:`.process_pair_db`.
Note that this may be called multiple times to process more
than one set of data files before computing profiles.
:param data_r1_path: path to R1 fragments
:param data_r2_path: path to matching R2 fragments.
"""
self.run._force_mask = force_mask
self.run.apply_config_restrictions()
self.force_mask = Mask(force_mask) if force_mask else None
use_quality = self.run._parse_quality
if not self.run.skip_database and not use_quality:
self.process_pair_db(self._memory_db_from_pairs(data_r1_path, data_r2_path))
else:
with FastFastqParser(data_r1_path, data_r2_path, use_quality) as parser:
if not self.run.pair_length:
self.run.pair_length = parser.pair_length()
self._process_pair_iter(parser.iterator(batch_size = 131072))
def __init__(self, masks):
if len(masks) == 2 and 'RRRY' in masks and 'YYYR' in masks:
self.match_mask = match_mask_optimized
else:
self.match_mask = self._match_mask
self.masks = []
for mask in masks:
self.masks.append(Mask(mask))
def main():
image = mycv2.loadImage(PARAMS["data"]["images"]["lena"])
mycv2.show(image)
mask = Mask(image)
plt.imshow(mask.mask, cmap="gray")
plt.show()
def testDefaults(self):
mask = Mask("Deku Mask", "Turns Link into a Deku shrub", True)
self.assertEqual(mask.name, "Deku Mask")
self.assertEqual(mask.properties, "Turns Link into a Deku shrub")
self.assertEqual(mask.isOn, False)
self.assertEqual(mask.acquired, False)
self.assertEqual(mask.cApplied, "none")
self.assertEqual(mask.inUse, False)
def get_indices(tid, si):
return set(
Mask.get_mask_stats_by_idx(
model.sess,
mask_id,
feed_dict={
X: model.trainXs[tid][si:si + batch_size_per_task],
is_training: True,
})["indices"])
def print_dead_quality():
orbit = 50000
m = Mask()
m.load_ascii_quality("qualities/"+str(orbit)+".txt")
print(np.where(m.mask))
print(np.sum(m.mask))
wlsmask = np.zeros(1024, dtype=np.bool)
wls_idx = np.array([143, 196, 219, 220, 222, 254, 338, 398, 399, 405, 406, 494, 502, 577, 601, 609, 624, 635, 678, 707, 779, 802, 868, 881, 902])
sun_idx = np.array([218, 219, 220, 609, 779, 883, 902, 1008])
wlsmask[wls_idx] = True
new_alive = m.mask.astype('i') - wlsmask
print("new alive = ", np.where(new_alive == 1), np.where(new_alive == 1)[0].size)
new_dead = wlsmask.astype('i') - m.mask
print("new dead = ", np.where(new_dead == 1), np.where(new_dead == 1)[0].size)
print(m.mask)
print(new_alive)
print(new_dead)
return
def CreateMask(
self,
imagedata=None,
name=None,
colour=None,
opacity=None,
threshold_range=None,
edition_threshold_range=None,
edited_points=None,
):
# TODO: mask system to new system.
future_mask = Mask()
future_mask.create_mask(self.matrix.shape)
if name:
future_mask.name = name
if colour:
future_mask.colour = colour
if opacity:
future_mask.opacity = opacity
if edition_threshold_range:
future_mask.edition_threshold_range = edition_threshold_range
if edited_points:
future_mask.edited_points = edited_points
if threshold_range:
future_mask.threshold_range = threshold_range
# insert new mask into project and retrieve its index
proj = Project()
index = proj.AddMask(future_mask)
future_mask.index = index
## update gui related to mask
Publisher.sendMessage(
"Add mask", (future_mask.index, future_mask.name, future_mask.threshold_range, future_mask.colour)
)
self.current_mask = future_mask
Publisher.sendMessage("Change mask selected", future_mask.index)
Publisher.sendMessage("Update slice viewer")
def crossover(parent1, parent2):
# Number of shapes from parents
n = random.randint(1, len(parent1.shapes))
m = random.randint(0, len(parent2.shapes))
# Adds parent genes to child
child = []
child += random.sample(parent1.shapes, n)
child += random.sample(parent2.shapes, m)
return Mask(child)
def init_population(popSize):
population = []
# Generates population
for _ in range(popSize):
# Makes individual and adds to population
indv = Mask()
population.append(indv)
return population
def run(self):
logging.info("starting data retention")
file_regex = [
'.*agencies.jsonl', '.*caregivers.jsonl', '.*care_logs.jsonl',
'.*clients.jsonl', '.*locations.jsonl', '.*shifts.jsonl',
'.*timezone_agencies.jsonl'
]
mask = Mask([
Condition("body.resource_url", file_mask)
for file_mask in file_regex
])
ResourceJoin(matched_callback=self.retain_data, mask=mask).join()
def __init__(self, dispatcher):
self.dispatcher = dispatcher
self.masks = {c: Mask(c, dispatcher) for c in COLORS}
self.entities = []
self.sides = [Side(name) for name in ['top', 'left', 'right']]
self.image = None
self.current_frame = None
self.tracked_entities = {}
self.width = None
self.height = None
self.collisions = None
self.balls = None
def __init__(self):
"""
Initializer
"""
super().__init__()
# Set the working directory (where we expect to find files) to the same
# directory this .py file is in. You can leave this out of your own
# code, but it is needed to easily run the examples using "python -m"
# as mentioned at the top of this program.
file_path = os.path.dirname(os.path.abspath(__file__))
os.chdir(file_path)
# Sprite lists
self.player = Player()
self.coin_list = None
self.wall_list = Solid_blocks().wall_list
self.player_list = self.player.player_list
self.player_sprite = self.player.player_sprite
self.player_health = self.player.player_health
self.brick_list = Destroyable_blocks().random_wall_list
self.virus = Virus_cells()
self.enemies = Virus_cells().virus_cells
self.mask = Mask()
self.mask_list = Mask().mask_list
self.power = None
self.walls_and_bricks = None
self.explosions_list = None
self.score = 0
self.mask_count = Mask().mask_count
self.physics_engine = None
self.volume = 0.4
self.background = None
self.background_music = None
self.width = constants.SCREEN_WIDTH
self.height = constants.SCREEN_HEIGHT
self.bullet_list = None
self.shotgun = False
self.mouse_clicks = 0
self.play_music = None
def build_model(self):
X_list = [tf.placeholder(tf.float32, [None, self.dims[0]], name="X_%d" % (t + 1))
for t in range(self.n_tasks)]
Y_list = [tf.placeholder(tf.float32, [None, self.n_classes], name="Y_%d" % (t + 1))
for t in range(self.n_tasks)]
is_training = tf.placeholder(tf.bool, name="is_training")
for t, (X, Y) in enumerate(zip(X_list, Y_list)):
bottom = X
excl_bottom = X if self.use_set_based_mask else None
for i in range(1, self.n_layers):
w = self.get_variable('layer%d' % i, 'weight', True)
b = self.get_variable('layer%d' % i, 'biases', True)
bottom = tf.nn.relu(tf.matmul(bottom, w) + b)
if self.use_set_based_mask:
bottom, residuals = Mask(
i - 1, self.mask_alpha[i - 1], self.mask_not_alpha[i - 1], mask_type=self.mask_type,
).get_masked_tensor(bottom, is_training, with_residuals=True)
mask = residuals["cond_mask"]
with tf.control_dependencies([mask]):
excl_bottom = tf.nn.relu(tf.matmul(excl_bottom, w) + b)
excl_bottom = Mask.get_exclusive_masked_tensor(excl_bottom, mask, is_training)
w = self.get_variable("layer%d" % self.n_layers, "weight_%d" % (t + 1), True)
b = self.get_variable("layer%d" % self.n_layers, "biases_%d" % (t + 1), True)
y = tf.matmul(bottom, w) + b
yhat = tf.nn.sigmoid(y)
self.yhat_list.append(yhat)
loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=y, labels=Y))
self.partial_loss_list.append(loss)
if self.use_set_based_mask:
excl_y = tf.matmul(excl_bottom, w) + b
excl_loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=excl_y, labels=Y))
self.excl_partial_loss_list.append(excl_loss)
return X_list, Y_list, is_training
def compare_combined_30_vs_32(orbit):
m30 = Mask()
m30.load_sdmf30_crit(orbit, "combined")
m32 = Mask()
m32.load_sdmf32_mask(orbit, "combinedFlag")
# for i in range(1024):
# print(i, m30.mask[i], m32.mask[i])
newdead = m30.get_new_dead(m32)
newalive = m30.get_new_alive(m32)
print("3.0->3.2 dead:", newdead.size, newdead)
print("3.0->3.2 alive:", newalive.size, newalive)
return
def Task2(self) -> int:
# Address masking
mem = {}
mask = None
for line in self.data:
if "mask" in line:
mask = Mask(line.split(" = ")[-1])
mask.mutate()
else:
value = int(line.split(" = ")[-1])
address = int(line.split("[")[-1].split("]")[0])
intermediate = mask.apply(address, value)
mem = {**mem, **intermediate}
result = 0
for value in mem.values():
if value != 0:
result += value
return result
def __init__(
self,
capture_link="https://imageserver.webcamera.pl/rec/hotel-senacki-krakow/latest.mp4",
box_min_area=800,
blur_size=15,
min_threshold=10,
frames_to_work=10,
average_alfa=0.1):
"""
:param capture_link: link to the stream (0 is for local camera)
:param box_min_area: minimal box area
:param blur_size: size of the blur, the bigger the larger boxes will be detected
:param min_threshold:
:param frames_to_work: frames to start detecting move
:param average_alfa: change of the background speed, between 0 and 1.
"""
self.background = bg.Background(frames_to_work, average_alfa)
self.captureStream = cv2.VideoCapture(capture_link)
self.box_min_area = box_min_area
self.blur_shape = (blur_size, blur_size)
self.min_threshold = min_threshold
self.state = States.Normal
self.mask = Mask()
self.defaults = (box_min_area, blur_size, min_threshold, average_alfa)
def print_nr_ppg(orbit):
dictwls = get_closest_state_exec(orbit, 61, "/SCIA/SDMF31/sdmf_extract_calib.h5", readoutMean=True)
orbit_range = dictwls["orbit_range"]
orbit = orbit_range[0]
print("closest wls orbit=", orbit)
m = Mask()
#
# first just print combined smoothmask
#
m.load_ascii(orbit)
print(np.where(m.mask))
fname = "/SCIA/SDMF30/sdmf_pixelmask.h5"
fid = h5py.File(fname, "r")
grpname = "orbitalMask/"
ds_orbits = fid[grpname+"orbitList"]
orbits = ds_orbits[:]
#print(orbits)
idx = orbit == orbits
if np.sum(idx) == 0:
raise Exception("orbit not in SDMF3.2 orbital mask")
i = np.argmax(idx)
print("i=",i)
ds_ppg = fid[grpname+"pixelGain"]
ppg_chan8 = ds_ppg[7*1024:,i]
print("ppg", np.sum(ppg_chan8), np.where(ppg_chan8))
return
def getWordPostion(self):
aux = []
width, height = self.matrix.getDimen()
matrix_copy = copy.copy(self.matrix)
mask = Mask('Mascaras/mask-dilation5x5.txt')
Manipulator.applyDilation(mask, matrix_copy)
lines = matrix_copy.getMatrixData()
positionLines = self.getInitialAndFinal(lines)
for l_initial, l_final in positionLines:
columns = Matrix.getTransposed(lines[l_initial:l_final + 1], width,
l_final - l_initial)
positionColumns = self.getInitialAndFinal(columns)
for c_initial, c_final in positionColumns:
aux += [((l_initial, c_initial), (l_final, c_final))]
return aux
def learn(self, layer0, goal):
'''
Train the network on a single batch.
:param layer0: the input batch
:param goal: the desired output of the final layer
'''
masks = {}
for i in self.maskPr:
layer = self.layers[i]
pR = self.maskPr[i]
maskI = Mask(layer, pR)
masks[i] = maskI
self.fire(layer0, masks)
self.train(goal, masks)
return self.layers[-1]
def CreateMask(self, imagedata=None, name=None, colour=None,
opacity=None, threshold_range=None,
edition_threshold_range = None,
edited_points=None):
future_mask = Mask()
if colour:
future_mask.colour = colour
if opacity:
future_mask.opacity = opacity
if edition_threshold_range:
future_mask.edition_threshold_range = edition_threshold_range
if edited_points:
future_mask.edited_points = edited_points
# this is not the first mask, so we will import data from old imagedata
if imagedata is None:
old_mask = self.current_mask
imagedata = old_mask.imagedata
future_mask.threshold_range = old_mask.threshold_range
if threshold_range:
future_mask.threshold_range = threshold_range
future_mask.imagedata = self.__create_mask_threshold(self.imagedata,
threshold_range)
else:
future_mask.imagedata = vtk.vtkImageData()
future_mask.imagedata.DeepCopy(imagedata)
future_mask.imagedata.Update()
# when this is not the first instance, user will have defined a name
if name is not None:
future_mask.name = name
if future_mask.is_shown:
self.blend_filter.SetOpacity(1, future_mask.opacity)
else:
self.blend_filter.SetOpacity(1, 0)
self.blend_filter.Update()
# insert new mask into project and retrieve its index
proj = Project()
index = proj.AddMask(future_mask)
future_mask.index = index
if threshold_range:
self.SetMaskThreshold(index, threshold_range)
future_mask.edited_points = {}
# update gui related to mask
ps.Publisher().sendMessage('Add mask',
(future_mask.index,
future_mask.name,
future_mask.threshold_range,
future_mask.colour))
self.current_mask = future_mask
ps.Publisher().sendMessage('Change mask selected', future_mask.index)
ps.Publisher().sendMessage('Update slice viewer')
self.calRegTerm()
self.updateThetaConstSize()
print "Interation index: %d, Costfunction value: %4f." % (
ii, self.error[ii])
if __name__ == '__main__':
from lens import LensList
from tcc import TCCList
from mask import Mask
from source import Source
import time
a = time.time()
m = Mask()
m.x_range = [-300.0, 300.0]
m.y_range = [-300.0, 300.0]
m.x_gridsize = 2.5
m.y_gridsize = 2.5
m.openGDS('./NanGateLibGDS/NOR2_X2.gds', 11, 0.3)
m.maskfft()
s = Source()
s.na = 1.35
s.maskxpitch = m.x_range[1] - m.x_range[0]
s.maskypitch = m.y_range[1] - m.y_range[0]
s.type = 'annular'
s.sigma_in = 0.7
s.sigma_out = 0.9
s.smooth_deta = 0.00
from pbm import Pbm
from mask import Mask
from manipulator import Manipulator
from text import Text
# Exemplo: Imagens/grupo_07_imagem_1_linhas_30_colunas_2_palavras_277.pbm
path = input('Digite o caminho do arquivo: ')
pbm = Pbm(path)
mask = Mask('Mascaras/mask-dilation.txt')
manipulator = Manipulator(pbm)
manipulator.applyMedian()
manipulator.applyDilation(mask)
text = Text(pbm)
text.markWords()
text.showInfo()
name = path.split('.')[0].split('/')[1]
pbm.save(name)
pretrain_file = "sskip.100.vectors"
tag_info_file = "tag.info"
#trn_file = "train.input.part"
#dev_file = "dev.input.part"
#tst_file = "test.input.part"
#pretrain_file = "sskip.100.vectors.part"
UNK = "<UNK>"
PADDING = "<PADDING>"
#############################################
## tags
tags = Tag(tag_info_file)
SOS = tags.SOS
EOS = tags.EOS
##############################################
##
mask_info = Mask(tags)
#############################################
trn_data = readfile(trn_file)
word_to_ix = {UNK:1, PADDING: 0}
lemma_to_ix = {UNK:1, PADDING: 0}
for sentence, _, lemmas, tags in trn_data:
for word in sentence:
if word not in word_to_ix:
word_to_ix[word] = len(word_to_ix)
for lemma in lemmas:
if lemma not in lemma_to_ix:
lemma_to_ix[lemma] = len(lemma_to_ix)
pretrain_to_ix = {UNK:1, PADDING: 0}
pretrain_embeddings = [ [0. for i in range(100)], [0. for i in range(100)]] # for UNK and PADDING
pretrain_data = readpretrain(pretrain_file)
self.AIList.append(self.AI)
self.RIList.append([])
for jj in self.doseList:
self.resist_t = self.resist_tRef*jj
self.calRI()
self.RIList[ii].append(self.RI)
if __name__ == "__main__":
from tcc import TCCList,TCC
from mask import Mask
from source import Source
from lens import LensList, Lens
mp = [ [[-1,6],[-1, 2],[1, 2],[1, 1],[6, 1],[6, 0],[0, 0],[0, 1],[-2, 1],[-2, 6],[-1, 6]], \
[[6, -1],[6, -2],[1, -2],[1, -3],[4, -3],[4, -6],[3, -6],[3, -4],[0, -4],[0, -1],[6, -1]] ]
m = Mask()
m.x_range = [-300.0,300.0]
m.y_range = [-400.0,300.0]
m.x_gridsize = 1.0
m.y_gridsize = 1.0
m.CD = 40
m.polygons = mp
m.poly2mask()
m.smooth()
m.maskfft()
"""nominal ILT setting"""
s = Source()
s.na = 1.35
s.maskxpitch = 600.0
s.maskypitch = 800.0
def main():
'''
Initialization
'''
start = time.time()
tar_4_corners_xy = None
ref_4_corners_xy = None
homoMat = np.load('image/save_H.npy')
shiftMat = np.load('image/save_Shift.npy')
ref_4_corners_xy = np.load('image/save_ref_4_corners_xy.npy')
tar_4_corners_xy = np.load('image/save_tar_4_corners_xy.npy')
warp_tar_img = cv2.imread('image/warped_target.png')
warp_ref_img = cv2.imread('image/warped_reference.png')
seam_mask = cv2.imread('image/seam_mask.png', cv2.IMREAD_GRAYSCALE)
mask = Mask(warp_tar_img, warp_ref_img)
ref_region_mask = cv2.imread('image/result_from_reference.png',
cv2.IMREAD_GRAYSCALE)
tar_region_mask = cv2.bitwise_and(cv2.bitwise_not(ref_region_mask),
mask.tar)
mask.tar_result = tar_region_mask
mask.ref_result = ref_region_mask
'''
Blend color
'''
CB = CBlender(warp_tar_img, warp_ref_img, seam_mask, tar_4_corners_xy,
ref_4_corners_xy, homoMat, shiftMat, mask)
devided_tar_nonoverlap_mask, devided_tar_edge_mask = CB.split_tar_nonoverlap_area_and_edge(
ref_4_corners_xy, tar_4_corners_xy, shiftMat, homoMat)
devided_tar_edge_mask_1 = np.copy(devided_tar_edge_mask)
devided_tar_edge_mask_1[devided_tar_edge_mask_1 == 127] = 0
devided_tar_edge_mask_2 = np.copy(devided_tar_edge_mask)
devided_tar_edge_mask_2[devided_tar_edge_mask_1 == 255] = 0
# refered_pixel_coordi_lst = np.vstack( (\
# np.array( np.nonzero(seam_mask) ).T, \
# np.array( np.where(devided_tar_edge_mask==255) ).T, \
# np.array( np.where(devided_tar_edge_mask==127) ).T \
# ))
slic = MaskedSLIC(warp_tar_img,
np.bitwise_and(mask.tar_result, mask.overlap),
region_size=20,
compactness=5)
seam_superpixel_idx_lst = np.array([
idx for idx, (rows, cols) in enumerate(slic.labels_position)
if np.sum(seam_mask[rows, cols]) > 0
])
tar_edge_1_superpixel_idx_lst = np.array([
idx for idx, (rows, cols) in enumerate(slic.labels_position)
if np.sum(devided_tar_edge_mask_1[rows, cols]) > 0
])
tar_edge_2_superpixel_idx_lst = np.array([
idx for idx, (rows, cols) in enumerate(slic.labels_position)
if np.sum(devided_tar_edge_mask_2[rows, cols]) > 0
])
refered_idx_lst = np.hstack(
(seam_superpixel_idx_lst, tar_edge_1_superpixel_idx_lst,
tar_edge_2_superpixel_idx_lst))
refered_pixel_coordi_lst = np.array([
(rows[len(rows) // 2], cols[len(rows) // 2])
for idx, (rows, cols) in enumerate(slic.labels_position)
if (idx != 0) and (np.isin(idx, refered_idx_lst))
])
fuck_this_shit = CB.blend_color(refered_pixel_coordi_lst,
np.bitwise_and(CB.mask.overlap,
CB.mask.tar_result),
sigma1=0.3,
sigma2=0.2)
cv2.imwrite('overlap_blending_only.png', fuck_this_shit)
CB.ref_img = fuck_this_shit
refered_pixel_coordi_lst = np.array([
(rows[len(rows) // 2], cols[len(rows) // 2])
for idx, (rows, cols) in enumerate(slic.labels_position)
if (idx != 0) and (np.isin(idx, tar_edge_1_superpixel_idx_lst))
])
mask = np.zeros(devided_tar_nonoverlap_mask.shape)
mask[devided_tar_nonoverlap_mask == 255] = 255
fuck_this_shit = CB.blend_color(refered_pixel_coordi_lst,
mask,
sigma1=0.1,
sigma2=0.05)
cv2.imwrite('nonoverlap_blending_only1.png', fuck_this_shit)
CB.ref_img = fuck_this_shit
refered_pixel_coordi_lst = np.array([
(rows[len(rows) // 2], cols[len(rows) // 2])
for idx, (rows, cols) in enumerate(slic.labels_position)
if (idx != 0) and (np.isin(idx, tar_edge_2_superpixel_idx_lst))
])
mask = np.zeros(devided_tar_nonoverlap_mask.shape)
mask[devided_tar_nonoverlap_mask == 127] = 255
fuck_this_shit = CB.blend_color(refered_pixel_coordi_lst,
mask,
sigma1=0.1,
sigma2=0.05)
cv2.imwrite('nonoverlap_blending_only2.png', fuck_this_shit)
for idx in tar_edge_2_superpixel_idx_lst:
fuck_this_shit[slic.labels_position[idx][0],
slic.labels_position[idx][1]] = np.random.randint(
0, 255, 3)
cv2.imwrite('araara.png', fuck_this_shit)
warp_ref_img[slic.contour_mask > 0] = (0, 255, 0)
warp_tar_img[slic.contour_mask > 0] = (0, 255, 0)
cv2.imwrite('target_image_slic.png', warp_ref_img)
cv2.imwrite('reference_image_slic.png', warp_tar_img)
print(f'time: {time.time() - start}')
def from_array(self, array):
if use_pygame:
if not size == self.get_size():
self = RectMaskCombo(self.rect, Mask((len(array), len(array[0]))))
Mask.from_array(self, array)
""" Created on Wen Apr 27 2016 @author: WenLv ([email protected]) """ from lens import LensList from tcc import TCCList from mask import Mask from source import Source from ilt import RobustILT import numpy as np import time m = Mask() m.x_gridsize = 2.5 m.y_gridsize = 2.5 m.openGDS('./NanGateLibGDS/NOR2_X2.gds',layername=11,boundary=0.3) m.maskfft() s = Source() s.na = 1.35 s.maskxpitch = m.x_range[1] - m.x_range[0] s.maskypitch = m.y_range[1] - m.y_range[0] s.type = 'annular' s.sigma_in = 0.7 s.sigma_out = 0.9 s.smooth_deta = 0.00 s.shiftAngle = 0
def compare_error_30_vs_32(orbit):
m30 = Mask()
m30_combined = Mask()
m30_combined.load_ascii(orbit)
m30.load_sdmf30_crit(orbit, "residual")
m32 = Mask()
m32.load_sdmf32_figure(orbit, "darkError")
newdead = m30.get_new_dead(m32)
newalive = m30.get_new_alive(m32)
print("new errory:", newdead.size, newdead)
print("new errory but already bad", np.sum(np.in1d(newdead, np.where(m30_combined.mask))))
print("real new errory", newdead[np.in1d(newdead, np.where(m30_combined.mask), invert=True)])
print("new not-errory:", newalive.size, newalive)
print("new not-errory but still bad", np.sum(np.in1d(newalive, np.where(m30_combined.mask))))
print("real new not-errory:", newalive[np.in1d(newalive, np.where(m30_combined.mask), invert=True)])
return
def collidepoint(self, point):
self.pos = self.rect.topleft
return Mask.collidepoint(self, point)
start = time.time()
H = None
shift = None
tar_4_corners_xy = None
ref_4_corners_xy = None
homoMat = np.load('image/save_H.npy')
shiftMat = np.load('image/save_Shift.npy')
ref_4_corners_xy = np.load('image/save_ref_4_corners_xy.npy')
tar_4_corners_xy = np.load('image/save_tar_4_corners_xy.npy')
warp_tar_img = cv2.imread('image/warped_target.png')
warp_ref_img = cv2.imread('image/warped_reference.png')
seam_mask = cv2.imread('image/seam_mask.png', cv2.IMREAD_GRAYSCALE)
mask = Mask(warp_tar_img, warp_ref_img)
ref_region_mask = cv2.imread('image/result_from_reference.png',
cv2.IMREAD_GRAYSCALE)
tar_region_mask = cv2.bitwise_and(cv2.bitwise_not(ref_region_mask), mask.tar)
mask.tar_result = tar_region_mask
mask.ref_result = ref_region_mask
CB = CBlender(warp_tar_img, warp_ref_img, seam_mask, tar_4_corners_xy,
ref_4_corners_xy, homoMat, shiftMat, mask)
devided_tar_nonoverlap_mask, devided_tar_edge_mask = CB.split_tar_nonoverlap_area_and_edge(
ref_4_corners_xy, tar_4_corners_xy, shiftMat, homoMat)
refered_pixel_coordi_lst = np.vstack( (\
np.array( np.nonzero(seam_mask) ).T, \
np.array( np.where(devided_tar_edge_mask==255) ).T, \
np.array( np.where(devided_tar_edge_mask==127) ).T \
def __init__(self, size, numbers_of_elements): Mask.__init__(self, size, numbers_of_elements) self.core = numbers_of_elements - size
self._retriever.stop()
def join(self):
self._retriever.join()
def __str__(self):
return str(self._mask)
def __repr__(self):
return self.__str__()
def blah(joiner):
print("blah received the following events:")
for event in joiner.mask().events():
print(event)
print("stopping the joiner!")
joiner.stop()
# e.g.
if __name__ == "__main__":
file_regex = ['.*agencies.jsonl',
'.*caregivers.jsonl',
'.*care_logs.jsonl',
'.*clients.jsonl',
'.*locations.jsonl',
'.*shifts.jsonl',
'.*timezone_agencies.jsonl']
mask = Mask([Condition("body.resource_url", file_mask) for file_mask in file_regex])
ResourceJoin(blah, mask).join()
class Processor:
def __init__( self ):
"""
Placeholder
"""
# increase system memory usage
os.environ['GDAL_CACHEMAX'] = '2048'
gdal.UseExceptions()
self._mask = Mask()
return
def getPansharpenImage( self, pathname ):
"""
Placeholder
"""
# get masked datasets
code = None
datasets = self.getMaskedDatasets ( os.path.dirname( pathname ) )
if len( datasets ) == 3:
# define pansharpen options
args = [ os.path.join( os.path.dirname( pathname ), 'B8_merge.tif' ) ] + datasets + [ pathname ]
options = [ '-co', 'PHOTOMETRIC=RGB', '-co', 'COMPRESS=DEFLATE', '-nodata', '0' ]
# manage execution of gdal pansharpen
out, err, code = execute( 'gdal_pansharpen.py', args + options )
return code
def getMaskedDatasets( self, path ):
"""
Placeholder
"""
# check qa image exists
datasets = []
qa_pathname = os.path.join( path, 'BQA_merge.tif' )
if os.path.exists( qa_pathname ):
arr = self._mask.generate( qa_pathname, out_pathname=os.path.join( path, 'mask.tif') )
if arr is not None:
# mask channel images
channels = [ 'B4', 'B3' , 'B2' ]
for channel in channels:
pathname = os.path.join( path, '{}_merge.tif'.format( channel ) )
out_pathname = pathname.replace( '_merge.tif', '_merge_mask.tif' )
# apply masking
arr = self._mask.apply( pathname, out_pathname=out_pathname )
if arr is not None:
datasets.append( out_pathname )
else:
# masking error
print ( 'Masking error - skipping: {}'.format ( qa_pathname ) )
else:
# no qa file
print ( 'File does not exist - skipping: {}'.format ( qa_pathname ) )
return datasets
def getImageChip( self, pathname, centroid, out_pathname, size=512, scale_min=2, scale_max=98 ):
"""
Placeholder
"""
# open pansharpen image
ds = gdal.Open( pathname )
if ds is not None:
# transform latlon centroid coordinates to image srs
coord_tx = self.getCoordinateTransform( ds )
x, y, z = coord_tx.TransformPoint( centroid[ 0 ], centroid[ 1 ] )
# convert image srs coordinates into pixel coordinates
geo = ds.GetGeoTransform()
pixel = ( round ( ( x - geo[ 0 ] ) / geo[ 1 ] ), round ( ( geo[ 3 ] - y ) / -geo[ 5 ] ) )
x_offset = ( pixel[ 0 ] - ( size / 2 ) )
y_offset = ( pixel[ 1 ] - ( size / 2 ) )
# check chip window is valid
if x_offset >= 0 and y_offset >= 0 and ( x_offset + size ) < ds.RasterXSize and ( y_offset + size ) < ds.RasterYSize:
# read in window
r = ds.GetRasterBand(1).ReadAsArray( xoff=x_offset, yoff=y_offset, win_xsize=size, win_ysize=size )
g = ds.GetRasterBand(2).ReadAsArray( xoff=x_offset, yoff=y_offset, win_xsize=size, win_ysize=size )
b = ds.GetRasterBand(3).ReadAsArray( xoff=x_offset, yoff=y_offset, win_xsize=size, win_ysize=size )
# validate goodness - number of black (masked) pixels
goodness = ( np.count_nonzero( r ) / ( size * size ) ) * 100.0
if goodness > 95.0:
# get scale factors to convert from 16bit to 8bit
scale = { 'r' : self.getScaleMinMax( r, ( scale_min, scale_max ) ),
'g' : self.getScaleMinMax( g, ( scale_min, scale_max ) ),
'b' : self.getScaleMinMax( b, ( scale_min, scale_max ) ) }
# compile translation options
options = [ '-of JPEG',
'-ot Byte',
'-srcwin {} {} {} {}'.format( x_offset, y_offset, size, size ),
'-scale_1 {} {}'.format ( scale[ 'r' ][ 0 ], scale[ 'r' ][ 1 ] ),
'-scale_2 {} {}'.format ( scale[ 'g' ][ 0 ], scale[ 'g' ][ 1 ] ),
'-scale_3 {} {}'.format ( scale[ 'b' ][ 0 ], scale[ 'b' ][ 1 ] ) ]
# create output directory if not exists
if not os.path.exists( os.path.dirname( out_pathname ) ):
os.makedirs( os.path.dirname( out_pathname ) )
# execute translation
options_str = ' '.join( options )
out = gdal.Translate( out_pathname, ds, options=options_str )
print ( 'Generated chip: {} {}'.format ( out_pathname, goodness ) )
else:
# failed qc check
print ( 'Chip failed QC: {} {}'.format( out_pathname, goodness ) )
else:
# report error
print ( 'Invalid chip window {} {} {} {} ({} {}): {}'.format( x_offset, y_offset, x_offset + size, y_offset + size,
ds.RasterXSize, ds.RasterYSize, pathname ) )
else:
# report error
print ( ' Unable to open dataset: {}'.format( pathname ) )
return
def getCoordinateTransform( self, ds ):
"""
Placeholder
"""
# get transformation from latlon to image srs
image = osr.SpatialReference( wkt=ds.GetProjection() )
aoi = osr.SpatialReference()
aoi.ImportFromEPSG( 4326 )
return osr.CoordinateTransformation( aoi, image )
# compute percentiles from cdf
def getScaleMinMax( self, arr, pcs ):
"""
Placeholder
"""
return np.percentile( arr[ arr != 0 ], pcs )
self.calRobustGrad()
self.calRegTerm()
self.updateThetaConstSize()
print "Interation index: %d, Costfunction value: %4f." %(ii, self.error[ii])
if __name__ == '__main__' :
from lens import LensList
from tcc import TCCList
from mask import Mask
from source import Source
import time
a = time.time()
m = Mask()
m.x_range = [-300.0,300.0]
m.y_range = [-300.0,300.0]
m.x_gridsize = 2.5
m.y_gridsize = 2.5
m.openGDS('./NanGateLibGDS/NOR2_X2.gds',11,0.3)
m.maskfft()
s = Source()
s.na = 1.35
s.maskxpitch = m.x_range[1] - m.x_range[0]
s.maskypitch = m.y_range[1] - m.y_range[0]
s.type = 'annular'
s.sigma_in = 0.7
s.sigma_out = 0.9
s.smooth_deta = 0.00
def colliderect(self, other):
self.pos = self.rect.topleft
return Mask.colliderect(self, other)
def save_mask_window(self):
""" Creates and saves the mask """
width = self.ui.lineEdit_width.text().replace(",", ".")
x = self.ui.lineEdit_x.text().replace(",", ".")
y = self.ui.lineEdit_y.text().replace(",", ".")
z = self.ui.lineEdit_z.text().replace(",", ".")
# If the coordinates and radius_width are valid numbers, then choose a given path and create the mask
if self.is_pos_number(width) and self.is_pos_number(x) and self.is_pos_number(y) and self.is_pos_number(z):
if float(width) == 0:
QMessageBox.warning(
self, "Warning", "You have entered invalid width/radius. Please enter numbers larger than zero.")
else:
# Convert coordinates and shape_size to voxels, using the affine matrix of the brain file
coordinate = np.array([float(self.ui.lineEdit_x.text()), float(self.ui.lineEdit_y.text()), float(self.ui.lineEdit_z.text())])
coordinate = apply_affine(np.linalg.inv(self.brain_file._affine), coordinate)
coordinate[0] = round(coordinate[0])
coordinate[1] = round(coordinate[1])
coordinate[2] = round(coordinate[2])
# If the coordinates together with radius_width are valid create a mask
# If not valid send an error message
if self.ui.comboBox_shape.currentText() == "Sphere":
voxel_width = (float(width)/self.brain_file._header.get_zooms()[0],
float(width)/self.brain_file._header.get_zooms()[1],
float(width)/self.brain_file._header.get_zooms()[2],)
else:
voxel_width = ((float(width)/2)/self.brain_file._header.get_zooms()[0],
(float(width)/2)/self.brain_file._header.get_zooms()[1],
(float(width)/2)/self.brain_file._header.get_zooms()[2])
# Check if coordinates is within brain data
# Generates parameters to and call Mask function
if coordinate[0] <= self.brain_file.get_data().shape[0]\
and coordinate[1] <= self.brain_file.get_data().shape[1]\
and coordinate[2] <= self.brain_file.get_data().shape[2]:
if (coordinate[0] + voxel_width[0]) <= self.brain_file.get_data().shape[0]\
and (coordinate[1] + voxel_width[1]) <= self.brain_file.get_data().shape[1]\
and (coordinate[2] + voxel_width[2]) <= self.brain_file.get_data().shape[2]\
and (coordinate[0] - voxel_width[0]) >= 0 and (coordinate[1] - voxel_width[1]) >= 0\
and (coordinate[2] - voxel_width[2]) >= 0:
file_name = QFileDialog.getSaveFileName(self, "Save file as nii", "", ".nii")
if file_name[0] == "":
return
path = file_name[0]+file_name[1]
shape = self.ui.comboBox_shape.currentText()
width = voxel_width
Mask(path, shape, coordinate, width, self.brain_file)
self.close()
self.parent().load_mask(path)
else:
QMessageBox.warning(
self, "Warning",
"You have entered invalid values in width. Please enter a value within the dimensions.")
else:
QMessageBox.warning(
self, "Warning",
"You have entered invalid values in coordinate. Please enter values within the dimensions.")
else:
QMessageBox.warning(
self, "Warning", "You have entered one or more invalid values. Please enter only numbers.")
