def edge_detection(model):
"""Image edge detection."""
# get data and name from current selected image
current_row = model.currentIndex().row()
source_vol = model.data(model.index(current_row), Qt.UserRole + 6)
source_name = model.data(model.index(current_row), Qt.DisplayRole)
# detect the edge
new_vol = imtool.multi_label_edge_detection(source_vol)
new_vol_name = "edge_" + source_name
# save result as a new image
model.addItem(new_vol, None, new_vol_name, model._data[0].get_header(), None, None, 255, "green")
return
def edge_detection(model):
"""Image edge detection."""
# get data and name from current selected image
current_row = model.currentIndex().row()
source_vol = model.data(model.index(current_row), Qt.UserRole + 6)
source_name = model.data(model.index(current_row), Qt.DisplayRole)
# detect the edge
new_vol = imtool.multi_label_edge_detection(source_vol)
new_vol_name = 'edge_' + source_name
# save result as a new image
model.addItem(new_vol, None, new_vol_name, model._data[0].get_header(),
None, None, 255, 'green')
return
def roi_projection(img, roi, dis_th, val_th, mode):
"""Roi projection."""
#st = time.time()
data = img
roi = imt.multi_label_edge_detection(roi)
result_mask = np.zeros(data.shape)
con_mask = np.zeros(data.shape)
tmp_mask = data > 0
roi_mask = roi > 0
img_indexs = np.transpose(tmp_mask.nonzero())
roi_indexs = np.transpose(roi_mask.nonzero())
spacedist = ds.cdist(img_indexs, roi_indexs, 'euclidean')
#print spacedist
mode = int(mode)
val_th = float(val_th)
# print mode,val_th
if mode == 0:
for dist in spacedist:
index = dist.argmin()
if dist.min() < dis_th:
coor = roi_indexs[index]
result_mask[coor[0]][coor[1]][coor[2]] = 1
elif mode == 1:
for dist in spacedist:
index = dist.argmin()
if dist.min() < dis_th:
coor = roi_indexs[index]
result_mask[coor[0]][coor[1]][
coor[2]] = result_mask[coor[0]][coor[1]][coor[2]] + 1
result_mask[result_mask < val_th] = 0
# print result_mask[result_mask.nonzero()]
elif mode == 2:
for i, dist in enumerate(spacedist):
index = dist.argmin()
if dist.min() < dis_th:
dcoor = img_indexs[i] # source coor
coor = roi_indexs[index] #target coor
result_mask[coor[0]][coor[1]][coor[2]] = result_mask[coor[0]][
coor[1]][coor[2]] + data[dcoor[0]][dcoor[1]][dcoor[2]]
con_mask[coor[0]][coor[1]][
coor[2]] = con_mask[coor[0]][coor[1]][coor[2]] + 1
result_mask = result_mask / (con_mask + 0.0000001)
result_mask[result_mask < val_th] = 0
# print result_mask[result_mask.nonzero()]
# print "Time : %s"%(time.time()-st)
return result_mask
def _edge_detection(self):
vol_name = str(self.out_edit.text())
if not vol_name:
self.out_edit.setFocus()
return
source_row = self.source_combo.currentIndex()
source_data = self._model.data(self._model.index(source_row),
Qt.UserRole + 5)
new_vol = imtool.multi_label_edge_detection(source_data)
self._model.addItem(new_vol, None, vol_name,
self._model._data[0].get_header(), None, None, 255,
'green')
self.done(0)
def roi_projection(img,roi,dis_th,val_th,mode):
"""Roi projection."""
#st = time.time()
data = img
roi = imt.multi_label_edge_detection(roi)
result_mask = np.zeros(data.shape)
con_mask = np.zeros(data.shape)
tmp_mask = data > 0
roi_mask = roi > 0
img_indexs = np.transpose(tmp_mask.nonzero())
roi_indexs = np.transpose(roi_mask.nonzero())
spacedist = ds.cdist(img_indexs,roi_indexs,'euclidean')
#print spacedist
mode = int(mode)
val_th = float(val_th)
# print mode,val_th
if mode ==0:
for dist in spacedist:
index = dist.argmin()
if dist.min()< dis_th:
coor = roi_indexs[index]
result_mask[coor[0]][coor[1]][coor[2]] = 1
elif mode == 1:
for dist in spacedist:
index = dist.argmin()
if dist.min()< dis_th:
coor = roi_indexs[index]
result_mask[coor[0]][coor[1]][coor[2]] = result_mask[coor[0]][coor[1]][coor[2]]+1
result_mask[ result_mask < val_th] = 0
# print result_mask[result_mask.nonzero()]
elif mode == 2:
for i,dist in enumerate(spacedist):
index = dist.argmin()
if dist.min()< dis_th:
dcoor = img_indexs[i] # source coor
coor = roi_indexs[index] #target coor
result_mask[coor[0]][coor[1]][coor[2]] = result_mask[coor[0]][coor[1]][coor[2]]+data[dcoor[0]][dcoor[1]][dcoor[2]]
con_mask[coor[0]][coor[1]][coor[2]] = con_mask[coor[0]][coor[1]][coor[2]]+1
result_mask = result_mask/(con_mask+0.0000001)
result_mask[ result_mask < val_th] = 0
# print result_mask[result_mask.nonzero()]
# print "Time : %s"%(time.time()-st)
return result_mask
def _edge_detection(self):
vol_name = str(self.out_edit.text())
if not vol_name:
self.out_edit.setFocus()
return
source_row = self.source_combo.currentIndex()
source_data = self._model.data(self._model.index(source_row),
Qt.UserRole + 5)
new_vol = imtool.multi_label_edge_detection(source_data)
self._model.addItem(new_vol,
None,
vol_name,
self._model._data[0].get_header(),
None, None, 255, 'green')
self.done(0)
def roi_to_gwmi(img,brain_wm,nth):
"""Transform the functional roi to wm.
Algorithm: find the nearest wm voxel.
"""
#st = time.time()
data = img
wmdata = imt.multi_label_edge_detection(brain_wm)
shape = data.shape
roi_ids = np.unique(data)
roi = roi_ids[1:]
roi = [int(i) for i in roi]
wmdata = wmdata!=0
result_mask = np.zeros(data.shape)
#print wmdata
#First, get the nonzero voxel index in image data.
#Here image data is a label volume.
#ROIs is in it
for roi_id in roi:
#print roi_id
tmp_mask = data==roi_id
#print tmp_mask
indexs = np.transpose(tmp_mask.nonzero())
#print indexs
#Second, find the nearest wm voxel for each indexs.
for coor in indexs:
#print coor
x = coor[0]
y = coor[1]
z = coor[2]
if wmdata[x,y,z]==1:
result_mask[x,y,z] = roi_id
else:
#find the nearest neighbor.
flag = False
radius = 1
mindist_voxel = []
mindist = 1000
while radius<nth:
neigh_list = get_neighbors(coor,radius,shape)
radius += 1
#find the nearest white matter voxel.
for n in neigh_list:
#print n
if wmdata[n[0],n[1],n[2]]==1:
flag = True
dist = np.sqrt((n[0]-x)**2+(n[1]-y)**2+(n[2]-z)**2)
# if the distance is smaller than tag, choose it to be nearest.
if dist < mindist:
mindist = dist
mindist_voxel = n
if flag:
break
#print mindist_voxel
if mindist_voxel!=[]:
result_mask[mindist_voxel[0],mindist_voxel[1],mindist_voxel[2]] = roi_id
"""
for roi_id in roi:
tmp_mask = result_mask==roi_id
roi_size = tmp_mask.sum()
print roi_id, roi_size
"""
#print "Time use: %s"%(time.time()-st)
return result_mask
def roi_to_gwmi_1(img,brain_wm):
""" Transform the functional roi to wm.
Algorithm: find the nearest wm voxel.
"""
#st = time.time()
neighbors = [[1,0,0],\
[-1,0,0],\
[0,1,0],\
[0,-1,0],\
[0,0,-1],\
[0,0,1],\
[1,1,0],\
[1,1,1],\
[1,1,-1],\
[0,1,1],\
[-1,1,1],\
[1,0,1],\
[1,-1,1],\
[-1,-1,0],\
[-1,-1,-1],\
[-1,-1,1],\
[0,-1,-1],\
[1,-1,-1],\
[-1,0,-1],\
[-1,1,-1],\
[0,1,-1],\
[0,-1,1],\
[1,0,-1],\
[1,-1,0],\
[-1,0,1],\
[-1,1,0]]
neighbors = np.array(neighbors)
data = img
wmdata = imt.multi_label_edge_detection(brain_wm)
shape = data.shape
roi_ids = np.unique(data)
roi = roi_ids[1:]
roi = [int(i) for i in roi]
wmdata = wmdata!=0
result_mask = np.zeros(shape)
#print wmdata
#First, get the nonzero voxel index in image data.
#Here image data is a label volume.
#ROIs is in it
for roi_id in roi:
tmp_mask = data==roi_id
#print tmp_mask
indexs = np.transpose(tmp_mask.nonzero())
#print indexs
#Second, find the nearest wm voxel for each indexs.
#print roi_id,indexs.shape[0]
for coor in indexs:
#x = coor[0]
#y = coor[1]
# z = coor[2]
if wmdata[coor[0],coor[1],coor[2]]==1:
pass
#result_mask[x,y,z] = roi_id
else:
#find the nearest neighbor.
neigh_list = [(ioff+coor).tolist() for ioff in neighbors]
#find the nearest white matter voxel.
for n in neigh_list:
if wmdata[n[0],n[1],n[2]]==1:
result_mask[n[0],n[1],n[2]] = roi_id
"""
for roi_id in roi:
tmp_mask = result_mask==roi_id
roi_size = tmp_mask.sum()
#print roi_id, roi_size
"""
#print "Time use: %s"%(time.time()-st)
return result_mask
def roi_to_gwmi(img,brain_wm,nth):
"""
Transform the functional roi to wm.
Algorithm: find the nearest wm voxel.
"""
#st = time.time()
data = img
wmdata = imt.multi_label_edge_detection(brain_wm)
shape = data.shape
roi_ids = np.unique(data)
roi = roi_ids[1:]
roi = [int(i) for i in roi]
wmdata = wmdata!=0
result_mask = np.zeros(data.shape)
#print wmdata
#First, get the nonzero voxel index in image data.
#Here image data is a label volume.
#ROIs is in it
for roi_id in roi:
#print roi_id
tmp_mask = data==roi_id
#print tmp_mask
indexs = np.transpose(tmp_mask.nonzero())
#print indexs
#Second, find the nearest wm voxel for each indexs.
for coor in indexs:
#print coor
x = coor[0]
y = coor[1]
z = coor[2]
if wmdata[x,y,z]==1:
result_mask[x,y,z] = roi_id
else:
#find the nearest neighbor.
flag = False
radius = 1
mindist_voxel = []
mindist = 1000
while radius<nth:
neigh_list = get_neighbors(coor,radius,shape)
radius += 1
#find the nearest white matter voxel.
for n in neigh_list:
#print n
if wmdata[n[0],n[1],n[2]]==1:
flag = True
dist = np.sqrt((n[0]-x)**2+(n[1]-y)**2+(n[2]-z)**2)
# if the distance is smaller than tag, choose it to be nearest.
if dist < mindist:
mindist = dist
mindist_voxel = n
if flag:
break
#print mindist_voxel
if mindist_voxel!=[]:
result_mask[mindist_voxel[0],mindist_voxel[1],mindist_voxel[2]] = roi_id
"""
for roi_id in roi:
tmp_mask = result_mask==roi_id
roi_size = tmp_mask.sum()
print roi_id, roi_size
"""
#print "Time use: %s"%(time.time()-st)
return result_mask
def roi_to_gwmi_1(img,brain_wm):
"""
Transform the functional roi to wm.
Algorithm: find the nearest wm voxel.
"""
#st = time.time()
neighbors = [[1,0,0],\
[-1,0,0],\
[0,1,0],\
[0,-1,0],\
[0,0,-1],\
[0,0,1],\
[1,1,0],\
[1,1,1],\
[1,1,-1],\
[0,1,1],\
[-1,1,1],\
[1,0,1],\
[1,-1,1],\
[-1,-1,0],\
[-1,-1,-1],\
[-1,-1,1],\
[0,-1,-1],\
[1,-1,-1],\
[-1,0,-1],\
[-1,1,-1],\
[0,1,-1],\
[0,-1,1],\
[1,0,-1],\
[1,-1,0],\
[-1,0,1],\
[-1,1,0]]
neighbors = np.array(neighbors)
data = img
wmdata = imt.multi_label_edge_detection(brain_wm)
shape = data.shape
roi_ids = np.unique(data)
roi = roi_ids[1:]
roi = [int(i) for i in roi]
wmdata = wmdata!=0
result_mask = np.zeros(shape)
#print wmdata
#First, get the nonzero voxel index in image data.
#Here image data is a label volume.
#ROIs is in it
for roi_id in roi:
tmp_mask = data==roi_id
#print tmp_mask
indexs = np.transpose(tmp_mask.nonzero())
#print indexs
#Second, find the nearest wm voxel for each indexs.
#print roi_id,indexs.shape[0]
for coor in indexs:
#x = coor[0]
#y = coor[1]
# z = coor[2]
if wmdata[coor[0],coor[1],coor[2]]==1:
pass
#result_mask[x,y,z] = roi_id
else:
#find the nearest neighbor.
neigh_list = [(ioff+coor).tolist() for ioff in neighbors]
#find the nearest white matter voxel.
for n in neigh_list:
if wmdata[n[0],n[1],n[2]]==1:
result_mask[n[0],n[1],n[2]] = roi_id
"""
for roi_id in roi:
tmp_mask = result_mask==roi_id
roi_size = tmp_mask.sum()
#print roi_id, roi_size
"""
#print "Time use: %s"%(time.time()-st)
return result_mask
