def updatePlots(self, event):
"""
@param event:
@return
"""
if self.plotter == None:
return
if self.filterSetAppy.get():
self.plotter.updateData(Filters.BackgroundSubtraction(
self.dataset))
else:
self.plotter.updateData(self.dataset)
self.spectra = PlotWindow.PlotWindow(self.plotWindow,
self.plotter.createSpectra())
self.lineHoriz = PlotWindow.PlotWindow(
self.plotWindow,
self.plotter.createYPointPlot(self.voltagePoint.get()))
self.lineVert = PlotWindow.PlotWindow(
self.plotWindow,
self.plotter.createXPointPlot(self.cyclePoint.get()))
self.plot = self._showPlot(self.lastSelectedPlot)
def ConnectedComponents(path,
data,
video,
is_slim,
outpath='local_path',
discriminant_thr=0.004,
denoising1_bkg=5,
dilation_bkg=6,
denoising1=10,
denoising2=0.81,
denoising2_sub=0.0078,
max_pool=2,
gauss_radius=2.0,
cc_thr=3000):
'''
Transformed into a function instead of only main
'''
run = data + '-' + video + '-'
output_path = nebbia_path + str(data) + '/' + str(
video) + '/trigger_thr' + str(discriminant_thr) + '/'
if outpath == 'local_path':
output_path = local_path + str(data) + '/' + str(
video) + '/trigger_thr' + str(discriminant_thr) + '/'
input_folder = path #+'frames/'
output_folder_mean = output_path + 'means_filtered_2den' + str(
denoising2) + '_gausrad' + str(gauss_radius) + '/'
output_folder_cc = output_path + 'cc_filtered_2den' + str(
denoising2) + '_gausrad' + str(gauss_radius) + '/'
if not os.path.exists(output_folder_mean):
os.makedirs(output_folder_mean)
if not os.path.exists(output_folder_cc):
os.makedirs(output_folder_cc)
#log_book
log_path = str(output_path) + 'log_book' + '_means_filtered_2den' + str(
denoising2) + '_gausrad' + str(gauss_radius) + '.txt'
total_number_cc = 0
signals = 0
start = time.time()
#log_text=open(log_path, "w")
#log_text.write("Copmuting total number of frames:\n")
#log_text.close()
n_frames = TriggerFunctions.TotalFrames(input_folder, 'outvid-' + run)
image_prototype = Image.open(
str(input_folder) + 'outvid-' + run + '001.png')
matrix_prototype = np.asarray(image_prototype.convert('L'))
print("Number of frames: " + str(n_frames) + '\n')
log_text = open(log_path, "w")
log_text.write('video ID' + str(run) + '\n')
log_text.write('denoising1_bkg ' + str(denoising1_bkg) + '\n')
log_text.write('dilation_bkg ' + str(dilation_bkg) + '\n')
log_text.write('denoising1 ' + str(denoising1) + '\n')
log_text.write('denoising2 ' + str(denoising2) + '\n')
log_text.write('denoising2_sub ' + str(denoising2_sub) + '\n')
log_text.write('max_pool ' + str(max_pool) + '\n')
log_text.write('gauss_radius ' + str(gauss_radius) + '\n')
log_text.write('cc_thr ' + str(cc_thr) + '\n')
log_text.write('discriminant_thr ' + str(discriminant_thr) + '\n')
log_text.write("Number of frames: " + str(n_frames) + '\n')
log_text.close()
#normalized mean and variance matrices
matrix_mean, matrix_var = TriggerFunctions.TotalMeanVar(
input_folder, matrix_prototype, 'outvid-' + run, n_frames)
if not is_slim:
plt.matshow(matrix_var)
plt.colorbar()
plt.savefig(str(output_path) + 'var_all.png')
plt.matshow(matrix_mean)
plt.colorbar()
plt.savefig(str(output_path) + 'mean_all.png')
#loop on frames
for file_raw in tqdm(os.listdir(input_folder)):
if not file_raw.endswith('.png'):
continue
if file_raw.endswith('-001.png'):
continue
raw_image = Image.open(str(input_folder) + file_raw)
matrix_raw = np.asarray(raw_image.convert('L'))
raw_file_name, dot, extension = file_raw.partition('.')
raw_file_name1, dot, raw_file_name2 = raw_file_name.partition('-')
raw_file_name3, dot, raw_file_name4 = raw_file_name2.partition('-')
raw_file_name5, dot, raw_file_name6 = raw_file_name4.partition('-')
if raw_file_name5 != video:
continue
#selecting via discriminant value
if not TriggerFunctions.ImageSelectingByDiscrimination(
matrix_raw / 255.,
matrix_mean,
matrix_var,
discriminant_thr,
verbose=True):
continue
print(raw_file_name5)
output_folder = output_path + 'pooled/' + str(raw_file_name2) + '/'
raw_video = raw_file_name1 + dot + raw_file_name3 + dot + raw_file_name5 + dot
signals += 1
log_text = open(log_path, "a")
log_text.write('Selected file number ' + str(signals) + ': ' +
str(file_raw) + '\n')
log_text.close()
print('Selected file number ' + str(signals) + ': ' + str(file_raw) +
'\n')
if not os.path.exists(output_folder):
os.makedirs(output_folder)
Filters.ReducingResolution(input_folder, file_raw, output_folder,
'raw_pooled.png', max_pool)
raw_pooled = Image.open(str(output_folder) + 'raw_pooled.png')
matrix_raw_pooled = np.asarray(raw_pooled.convert('L'))
matrix_mean_pooled = np.zeros_like(matrix_raw_pooled)
#subtracting backgrounds procedure
#counting the number of background frames
n_backgrounds = 0
start_partial = time.time()
for file in os.listdir(input_folder):
if (file.startswith(raw_video)):
if (file.endswith('-001.png') or file == file_raw):
continue
file_name, dot, extension = file.partition('.')
file_name1, dot, file_name2 = file_name.partition('-')
file_name3, dot, file_name4 = file_name2.partition('-')
file_name5, dot, file_name6 = file_name4.partition('-')
number = int(file_name6)
if (number % 2):
continue
n_backgrounds += 1
log_text = open(log_path, "a")
log_text.write('Total number of background images processed: ' +
str(n_backgrounds) + '\n')
log_text.close()
print('Total number of background images processed: ' +
str(n_backgrounds) + '\n')
for file in os.listdir(input_folder):
if (file.startswith(raw_video)):
if (file.endswith('-001.png') or file == file_raw):
continue
file_name, dot, extension = file.partition('.')
file_name1, dot, file_name2 = file_name.partition('-')
file_name3, dot, file_name4 = file_name2.partition('-')
file_name5, dot, file_name6 = file_name4.partition('-')
number = int(file_name6)
if (number % 2):
continue
#background manipulation before subtracting
Filters.Denoising1(input_folder, file, output_folder,
'bkg_denoising1.png', denoising1_bkg)
Filters.Dilation(output_folder, 'bkg_denoising1.png',
output_folder, 'bkg_dilationed.png',
dilation_bkg)
#raw_image manipulation before subtracting
Filters.Denoising1(input_folder, file_raw, output_folder,
'raw_denoising1.png', denoising1)
#background subtraction
Filters.BackgroundSubtraction(output_folder,
'raw_denoising1.png',
'bkg_dilationed.png',
output_folder, 'subtracted.png')
#manipulation after subtracting
Filters.Denoising2(output_folder, 'subtracted.png',
output_folder, 'sub_denoising2.png',
denoising2_sub)
Filters.ReducingResolution(output_folder, 'sub_denoising2.png',
output_folder, 'pool_' + str(file),
max_pool)
image_pooled = Image.open(
str(output_folder) + 'pool_' + str(file))
matrix_pooled = np.asarray(image_pooled.convert('L'))
matrix_mean_pooled = np.add(matrix_mean_pooled * 1.,
matrix_pooled * 1. / n_backgrounds)
image_mean = Image.fromarray(np.uint8(matrix_mean_pooled))
image_mean.save(
str(output_folder_mean) + 'mean_' + str(raw_file_name2) + '.png')
Filters.Denoising2(output_folder_mean,
'mean_' + str(raw_file_name2) + '.png',
output_folder_mean,
'mean_' + str(raw_file_name2) + '_den2.png',
denoising2)
number_cc = Filters.Labeling(output_folder_mean,
'mean_' + str(raw_file_name2) +
'_den2.png',
output_folder_cc,
'labeled_' + str(raw_file_name2) + '.png',
gauss_radius,
cc_thr,
verbose=True)
total_number_cc = total_number_cc + number_cc
stop_partial = time.time()
delta_partial = stop_partial - start_partial
delta_integrated = stop_partial - start
log_text = open(log_path, "a")
log_text.write("Time required for one frame to be analyzed: " +
str(delta_partial) + ' sec \n' +
"Integrated time from the beginning: " +
str(delta_integrated) + ' sec \n')
log_text.close()
print("Time required for one frame to be analyzed: " +
str(delta_partial) + ' sec \n' +
"Integrated time from the beginning: " + str(delta_integrated) +
' sec \n')
if is_slim:
os.system('rm -r ' + output_folder_mean + '../pooled')
os.system('rm -r ' + output_folder_mean)
os.system('rm ' + output_folder_cc + 'label*')
stop = time.time()
delta = stop - start
log_text = open(log_path, "a")
log_text.write("Total time for the execution: " + str(delta) + ' sec \n' +
"Total number of frames triggered for the analysis: " +
str(signals) + '\n' +
"Total number of cc collected in the video: " +
str(total_number_cc) + '\n')
log_text.close()
print("Total time for the execution: " + str(delta) + ' sec \n' +
"Total number of frames triggered for the analysis: " +
str(signals) + '\n' + "Total number of cc collected in the video: " +
str(total_number_cc) + '\n')
file_name3, dot, file_name4 = file_name2.partition('-')
file_name5, dot, file_name6 = file_name4.partition('-')
number = int(file_name6)
if (number % 2):
continue
#background manipulation before subtracting
Filters.Denoising1(input_folder, file, output_folder,
'bkg_denoising1.png', denoising1_bkg)
Filters.Dilation(output_folder, 'bkg_denoising1.png',
output_folder, 'bkg_dilationed.png', dilation_bkg)
#raw_image manipulation before subtracting
Filters.Denoising1(input_folder, file_raw, output_folder,
'raw_denoising1.png', denoising1)
#background subtraction
Filters.BackgroundSubtraction(output_folder, 'raw_denoising1.png',
'bkg_dilationed.png', output_folder,
'subtracted.png')
#manipulation after subtracting
Filters.Denoising2(output_folder, 'subtracted.png', output_folder,
'sub_denoising2.png', denoising2_sub)
Filters.ReducingResolution(output_folder, 'sub_denoising2.png',
output_folder, 'pool_' + str(file),
max_pool)
image_pooled = Image.open(str(output_folder) + 'pool_' + str(file))
matrix_pooled = np.asarray(image_pooled.convert('L'))
matrix_mean_pooled = np.add(matrix_mean_pooled * 1.,
matrix_pooled * 1. / n_backgrounds)
image_mean = Image.fromarray(np.uint8(matrix_mean_pooled))
image_mean.save(
str(output_folder_mean) + 'mean_' + str(raw_file_name2) + '.png')
def VideoAnalysis(path_out, video_name, frames_folder, traces_folder, log_path,
stat_path):
print('STARTING VIDEO ANALYSIS:\n')
#temporary folder to save filtered image before labeling
output_folder_mean = path_out + 'toLabel/'
output_folder = path_out + 'pooled/'
if not os.path.exists(output_folder_mean):
os.makedirs(output_folder_mean)
if not os.path.exists(output_folder):
os.makedirs(output_folder)
#setup parameters
signals = 0
total_traces = 0
start = time.time()
n_frames = TriggerFunctions.TotalFrames(frames_folder)
image_prototype = Image.open(str(frames_folder) + video_name + '-1.png')
matrix_prototype = np.asarray(image_prototype.convert('L'))
print("Number of frames: " + str(n_frames) + '\n')
log_text = open(log_path, "a")
log_text.write("Number of frames analyzed: " + str(n_frames) + '\n')
log_text.close()
#Normalized mean and variance matrices
matrix_mean, matrix_var = TriggerFunctions.TotalMeanVar(
frames_folder, matrix_prototype, n_frames)
#Saving mean and variance matrices
plt.matshow(matrix_var)
plt.colorbar()
plt.savefig(str(stat_path) + 'var_' + str(video_name) + '.png')
plt.matshow(matrix_mean)
plt.colorbar()
plt.savefig(str(stat_path) + 'mean_' + str(video_name) + '.png')
#LOOP ON FRAMES
for file_raw in tqdm(os.listdir(frames_folder)):
if not file_raw.endswith('.png'):
continue
if (file_raw.endswith('-1.png') or file_raw.endswith('-2.png')
or file_raw.endswith('-3.png') or file_raw.endswith('-4.png')):
continue
raw_image = Image.open(str(frames_folder) + file_raw)
matrix_raw = np.asarray(raw_image.convert('L'))
#Selecting relevant frames via discriminant value
if not TriggerFunctions.ImageSelectingByDiscrimination(
matrix_raw / 255.,
matrix_mean,
matrix_var,
discriminant_thr,
verbose=True):
continue
signals += 1
log_text = open(log_path, "a")
log_text.write('\nSelected file number ' + str(signals) + ': ' +
str(file_raw) + '\n')
log_text.close()
print('\nSelected file number ' + str(signals) + ': ' + str(file_raw) +
'\n')
raw_file_name, dot, formato = file_raw.partition('.')
temp_file_name = raw_file_name.split('-')
number_frame = int(temp_file_name[-1])
Filters.ReducingResolution(frames_folder, file_raw, output_folder,
'raw_pooled.png', max_pool)
raw_pooled = Image.open(str(output_folder) + 'raw_pooled.png')
matrix_raw_pooled = np.asarray(raw_pooled.convert('L'))
matrix_mean_pooled = np.zeros_like(matrix_raw_pooled)
#SUBTRACTING BACKGROUNDS PROCEDURE
#counting the number of background frames
n_backgrounds = 0
start_partial = time.time()
for file in os.listdir(frames_folder):
if (file.startswith(video_name)):
if (file.endswith('-1.png') or file.endswith('-2.png')
or file.endswith('-3.png') or file.endswith('-4.png')
or file == file_raw):
continue
file_name, dot, extension = file.partition('.')
file_name1 = file_name.split('-')
number = int(file_name1[-1])
if (number % 3):
continue
n_backgrounds += 1
log_text = open(log_path, "a")
log_text.write('Total number of background images processed: ' +
str(n_backgrounds) + '\n')
log_text.close()
print('Total number of background images processed: ' +
str(n_backgrounds) + '\n')
for file in os.listdir(frames_folder):
if (file.startswith(video_name)):
if (file.endswith('-1.png') or file.endswith('-2.png')
or file.endswith('-3.png') or file.endswith('-4.png')
or file == file_raw):
continue
file_name, dot, extension = file.partition('.')
file_name1 = file_name.split('-')
number = int(file_name1[-1])
if (number % 3):
continue
#background manipulation before subtracting
Filters.Denoising1(frames_folder, file, output_folder,
'bkg_denoising1.png', denoising1_bkg)
Filters.Dilation(output_folder, 'bkg_denoising1.png',
output_folder, 'bkg_dilationed.png',
dilation_bkg)
#raw_image manipulation before subtracting
Filters.Denoising1(frames_folder, file_raw, output_folder,
'raw_denoising1.png', denoising1)
#background subtraction
Filters.BackgroundSubtraction(output_folder,
'raw_denoising1.png',
'bkg_dilationed.png',
output_folder, 'subtracted.png')
#manipulation after subtracting
Filters.Denoising2(output_folder, 'subtracted.png',
output_folder, 'sub_denoising2.png',
denoising2_sub)
Filters.ReducingResolution(output_folder, 'sub_denoising2.png',
output_folder, 'pool_' + str(file),
max_pool)
image_pooled = Image.open(
str(output_folder) + 'pool_' + str(file))
matrix_pooled = np.asarray(image_pooled.convert('L'))
matrix_mean_pooled = np.add(matrix_mean_pooled * 1.,
matrix_pooled * 1. / n_backgrounds)
#meaning subtracted images
image_mean = Image.fromarray(np.uint8(matrix_mean_pooled))
image_mean.save(
str(output_folder_mean) + 'mean_' + str(raw_file_name) + '.png')
#binarization
Filters.Denoising2(output_folder_mean,
'mean_' + str(raw_file_name) + '.png',
output_folder_mean,
'mean_' + str(raw_file_name) + '_den2.png',
denoising2)
#LABELING
temp_folder_1 = path_out + 'temp1/'
temp_folder_2 = path_out + 'temp2/'
if os.path.exists(temp_folder_1):
shutil.rmtree(temp_folder_1)
if os.path.exists(temp_folder_2):
shutil.rmtree(temp_folder_2)
os.makedirs(temp_folder_1)
os.makedirs(temp_folder_2)
#first labeling
n1 = Filters.HiLoLabeling(output_folder_mean,
'mean_' + str(raw_file_name) + '_den2.png',
traces_folder, temp_folder_1, gauss_radius_1,
loCcThr_1, hiCcThr_1)
#second labeling
n2 = 0
for temp_cc in os.listdir(temp_folder_1):
q = Filters.HiLoLabeling(temp_folder_1, temp_cc, traces_folder,
temp_folder_2, gauss_radius_2, loCcThr_2,
hiCcThr_2)
n2 += q
#deleting temporary folders
shutil.rmtree(temp_folder_1)
shutil.rmtree(temp_folder_2)
total_cc = n1 + n2 #+n3
total_traces += total_cc
log_text = open(log_path, "a")
'''
log_text.write("Number of traces finded: "+str(total_cc)+'\n')
log_text.write("Time required for one frame to be analyzed: "+str(delta_partial)+' sec \n'+
"Integrated time from the beginning: "+str(delta_integrated)+' sec \n')
log_text.close()
print("Number of traces finded: "+str(total_cc)+'\n')
print("Time required for one frame to be analyzed: "+str(delta_partial)+' sec \n'+
"Integrated time from the beginning: "+str(delta_integrated)+' sec \n')
'''
stop = time.time()
delta = stop - start
log_text = open(log_path, "a")
log_text.write("\nTotal time for the execution: " + str(delta) +
' sec\n\n' +
"Total number of frames selected for the analysis: " +
str(signals) + '\n')
log_text.write("Total traces:" + str(total_traces) + str('\n\n\n'))
log_text.close()
print("Total time for the execution: " + str(delta) + ' sec \n\n' +
"Total number of frames selected for the analysis: " + str(signals) +
'\n')
print("Total traces:" + str(total_traces) + str('\n\n\n'))
#deleting temporary folders
shutil.rmtree(output_folder_mean)
shutil.rmtree(output_folder)
