updated for lwir16 (b88b2585) · Commits · Elphel / python3-imagej-tiff

imagej_tiff.py

+423 −433

Original line number	Original line	Diff line number	Diff line
	@@ -80,7 +80,7 @@ def get_tile_images(image, width=8, height=8):
	#from libtiff import TIFF		#from libtiff import TIFF
	'''		'''
	Description:		Description:
	Reads a tiff files with multiple layers that were saved by imagej		Reads TIFF files with multiple layers that were saved by imagej
	Methods:		Methods:
	.getstack(items=[])		.getstack(items=[])
	returns np.array, layers are stacked along depth - think of RGB channels		returns np.array, layers are stacked along depth - think of RGB channels
	@@ -95,12 +95,10 @@ Examples:

	'''		'''
	class imagej_tiff:		class imagej_tiff:

	# imagej stores labels lengths in this tag		# imagej stores labels lengths in this tag
	__TIFF_TAG_LABELS_LENGTHS = 50838		__TIFF_TAG_LABELS_LENGTHS = 50838
	# imagej stores labels conents in this tag		# imagej stores labels conents in this tag
	__TIFF_TAG_LABELS_STRINGS = 50839		__TIFF_TAG_LABELS_STRINGS = 50839

	# init		# init
	def __init__(self,filename, layers = None, tile_list = None):		def __init__(self,filename, layers = None, tile_list = None):
	# file name		# file name
	@@ -133,10 +131,7 @@ class imagej_tiff:
	tif.seek(i)		tif.seek(i)
	a = np.array(tif)		a = np.array(tif)
	a = np.reshape(a,(a.shape[0],a.shape[1],1))		a = np.reshape(a,(a.shape[0],a.shape[1],1))

	#a = a[:,:,np.newaxis]		#a = a[:,:,np.newaxis]

	# scale for 8-bits
	# exclude layer named 'other'		# exclude layer named 'other'
	if self.bpp==8:		if self.bpp==8:
	_min = self.data_min		_min = self.data_min
	@@ -151,9 +146,7 @@ class imagej_tiff:
	a[a==0]=np.nan		a[a==0]=np.nan
	a = a.astype(float)		a = a.astype(float)
	if self.labels[i]!='other':		if self.labels[i]!='other':
	# a[a==0]=np.nan
	a = (_max-_min)*(a-_MIN)/(_MAX-_MIN)+_min		a = (_max-_min)*(a-_MIN)/(_MAX-_MIN)+_min
	# init
	if i==0:		if i==0:
	self.image = a		self.image = a
	# stack along depth (think of RGB channels)		# stack along depth (think of RGB channels)
	@@ -213,10 +206,15 @@ class imagej_tiff:
	self.target_disparity = tiles_other[...,0]		self.target_disparity = tiles_other[...,0]
	self.gt_ds = tiles_other[...,1:3]		self.gt_ds = tiles_other[...,1:3]
	pass		pass

	# init done, close the image		# init done, close the image
	tif.close()		if (self.props['VERSION']== 2.0):
			# self.tileH = self.image.shape[0]//self.props['tileStepY']
			# self.tileW = self.image.shape[1]//self.props['tileStepX']
			self.tileH = self.props['tileStepY']
			self.tileW = self.props['tileStepX']
			pass

			tif.close()
	# label == tiff layer name		# label == tiff layer name
	def getvalues(self,label=""):		def getvalues(self,label=""):
	l = self.getstack([label],shape_as_tiles=True)		l = self.getstack([label],shape_as_tiles=True)
	@@ -283,11 +281,8 @@ class imagej_tiff:

	return res		return res




	# get ordered stack of images by provided items		# get ordered stack of images by provided items
	# by index or label name		# by index or label name. Divides into [self.tileH][self.tileW] tiles
	def getstack(self,items=[],shape_as_tiles=False):		def getstack(self,items=[],shape_as_tiles=False):
	a = ()		a = ()
	if len(items)==0:		if len(items)==0:
	@@ -301,17 +296,30 @@ class imagej_tiff:
	a += (self.image[:,:,j],)		a += (self.image[:,:,j],)
	# stack along depth		# stack along depth
	b = np.stack(a,axis=2)		b = np.stack(a,axis=2)

	if shape_as_tiles:		if shape_as_tiles:
	b = get_tile_images(b,self.tileW,self.tileH)		b = get_tile_images(b,self.tileW,self.tileH)

	return b		return b

			def trimStack (self, stack, radius = 0):
			if (radius == 0):
			radius=self.props['corrRadius']
			corr_side = 2*radius+1
			return stack[:,:,:,:corr_side,:corr_side]
	# get np.array of a channel		# get np.array of a channel
	# * do not handle out of bounds		# * do not handle out of bounds
	def channel(self,index):		def channel(self,index):
	return self.image[:,:,index]		return self.image[:,:,index]

			def getCorrsMeta(self,items=[]):
			stack0 = self.getstack(items,shape_as_tiles=True)
			stack = np.moveaxis(stack0, 4, 0) # slices - first index
			radius=self.props['corrRadius']
			num_meta=self.props['numMeta']
			corr_side = 2*radius+1
			corr_tiles = stack[:,:,:,:corr_side,:corr_side]
			meta = stack[:,:,:,-1,:num_meta]
			return corr_tiles, meta/self.props['tileMetaScale']

	# display images by index or label		# display images by index or label
	def show_images(self,items=[]):		def show_images(self,items=[]):
	@@ -332,7 +340,6 @@ class imagej_tiff:

	# display single image		# display single image
	def show_image(self,index):		def show_image(self,index):

	# display using matplotlib		# display using matplotlib

	t = self.image[:,:,index]		t = self.image[:,:,index]
	@@ -359,7 +366,6 @@ class imagej_tiff:
	infos = []		infos = []
	for info in self.infos:		for info in self.infos:
	infos.append(ET.fromstring(info))		infos.append(ET.fromstring(info))

	self.infos = infos		self.infos = infos

	# specifics		# specifics
	@@ -372,16 +378,46 @@ class imagej_tiff:
	pd[child.tag] = child.text		pd[child.tag] = child.text

	self.props = pd		self.props = pd
			file_version = float(self.props['VERSION'])
	# tiles are squares		if (file_version < 2.0):
			# tiles are squares (older version
	self.tileW = int(self.props['tileWidth'])		self.tileW = int(self.props['tileWidth'])
	self.tileH = int(self.props['tileWidth'])		self.tileH = int(self.props['tileWidth'])
	self.data_min = float(self.props['data_min'])		self.data_min = float(self.props['data_min'])
	self.data_max = float(self.props['data_max'])		self.data_max = float(self.props['data_max'])
			else:
			floats=['dispOffsetLow','tileMetaScale','disparity_low','dispOffset',
			'fatZero','disparity_pwr','VERSION','dispOffsetHigh',
			'disparity_high']
			ints = ['metaGTConfidence','tileMetaSlice','indexReference','metaLastDiff',
			'metaGTDisparity','metaFracValid', 'numScenes','metaTargetDisparity',
			'disparity_steps','tileStepX', 'tileStepY', "corrRadius","numMeta"]
			bools=['randomize_offsets']
			for key in pd:
			val = pd[key]
			if key in bools:
			if (val == '1') or (val == 'true') or (val == 'True'):
			pd[key] = 1
			else:
			pd[key] = 0
			pass
			elif key in ints:
			pd[key] = int(pd[key])
			elif key in floats:
			pd[key] = float(pd[key])
			try:
			pd['corrRadius'] = pd['corrRadius'] # not yet exists
			except:
			pd['corrRadius'] = 7
			try:
			pd['numMeta'] = pd['numMeta'] # not yet exists
			except:
			pd['numMeta'] = 6

			pass

	# makes arrays of labels (strings) and unparsed xml infos		# makes arrays of labels (strings) and unparsed xml infos
	def __split_labels(self,n,tag):		def __split_labels(self,n,tag):

	# list		# list
	tag_lens = tag[self.__TIFF_TAG_LABELS_LENGTHS]		tag_lens = tag[self.__TIFF_TAG_LABELS_LENGTHS]
	# string		# string
	@@ -405,21 +441,13 @@ class imagej_tiff:
	skip -= 1		skip -= 1
	tag_labels = tag_labels[l:]		tag_labels = tag_labels[l:]


	#MAIN		#MAIN
	if __name__ == "__main__":		if __name__ == "__main__":

	try:		try:
	fname = sys.argv[1]		fname = sys.argv[1]
	except IndexError:		except IndexError:
	fname = "/mnt/dde6f983-d149-435e-b4a2-88749245cc6c/home/eyesis/x3d_data/data_sets/train/1527182807_896892/v02/ml/1527182807_896892-ML_DATA-08B-O-FZ0.05-OFFS0.40000.tiff"		#fname = "/mnt/dde6f983-d149-435e-b4a2-88749245cc6c/home/eyesis/x3d_data/data_sets/train/1527182807_896892/v02/ml/1527182807_896892-ML_DATA-08B-O-FZ0.05-OFFS0.40000.tiff"
	# fname = "1521849031_093189-ML_DATA-32B-O-OFFS1.0.tiff"		fname = "/home/elphel/lwir16-proc/proc1/models/1626032208_613623/v01/ml32/1626032208_613623-ML-AUX-RND-DOFFS-5.000.tiff"
	# fname = "1521849031_093189-ML_DATA-08B-O-OFFS1.0.tiff"

	#fname = "1521849031_093189-DISP_MAP-D0.0-46.tif"
	#fname = "1526905735_662795-ML_DATA-08B-AIOTD-OFFS2.0.tiff"
	#fname = "test.tiff"

	print(bcolors.BOLDWHITE+"time: "+str(time.time())+bcolors.ENDC)		print(bcolors.BOLDWHITE+"time: "+str(time.time())+bcolors.ENDC)

	ijt = imagej_tiff(fname)		ijt = imagej_tiff(fname)
	@@ -432,32 +460,28 @@ if __name__ == "__main__":
	rough_string = ET.tostring(ijt.infos[0], "utf-8")		rough_string = ET.tostring(ijt.infos[0], "utf-8")
	reparsed = minidom.parseString(rough_string)		reparsed = minidom.parseString(rough_string)
	print(reparsed.toprettyxml(indent="\t"))		print(reparsed.toprettyxml(indent="\t"))

	#print(ijt.props)

	# needed properties:		# needed properties:
	print("Tiles shape: "+str(ijt.tileW)+"x"+str(ijt.tileH))		print("Tiles shape: "+str(ijt.tileW)+"x"+str(ijt.tileH))
			try:
	print("Data min: "+str(ijt.data_min))		print("Data min: "+str(ijt.data_min))
	print("Data max: "+str(ijt.data_max))		print("Data max: "+str(ijt.data_max))
			except:
			print("Data min/max are not provided")

	print(ijt.image.shape)		print(ijt.image.shape)
			# tiles,tiles_meta = ijt.getCorrsMeta(['0-1','1-2','2-3','3-4'])
			tiles,tiles_meta = ijt.getCorrsMeta([])
			print("Corr stack shape: "+str(tiles.shape))
			print("Meta stack shape: "+str(tiles_meta.shape))

	# layer order: ['diagm-pair', 'diago-pair', 'hor-pairs', 'vert-pairs', 'other']		exit (0)
	# now split this into tiles:		#
			# each tile's disparity:
	#tiles = get_tile_images(ijt.image,ijt.tileW,ijt.tileH)
	#print(tiles.shape)

	tiles = ijt.getstack(['diagm-pair','diago-pair','hor-pairs','vert-pairs'],shape_as_tiles=True)
	print("Stack of images shape: "+str(tiles.shape))

	print(bcolors.BOLDWHITE+"time: "+str(time.time())+bcolors.ENDC)
	# provide layer name		# provide layer name
	values = ijt.getvalues(label='other')		values = ijt.getvalues(label='other')
	print("Stack of values shape: "+str(values.shape))		print("Stack of values shape: "+str(values.shape))

	# each tile's disparity:

	fig = plt.figure()		fig = plt.figure()
	fig.suptitle("Estimated Disparity")		fig.suptitle("Estimated Disparity")
	plt.imshow(values[:,:,0])		plt.imshow(values[:,:,0])
	@@ -474,40 +498,6 @@ if __name__ == "__main__":
	plt.colorbar()		plt.colorbar()

	print(bcolors.BOLDWHITE+"time: "+str(time.time())+bcolors.ENDC)		print(bcolors.BOLDWHITE+"time: "+str(time.time())+bcolors.ENDC)
	#print(values)

	#print(value_tiles[131,162].flatten())
	#print(np.ravel(value_tiles[131,162]))

	#values = np.empty((vt.shape[0],vt.shape[1],3))

	#for i in range(values.shape[0]):
	# for j in range(values.shape[1]):
	# values[i,j,0] = get_v1()


	#print(tiles[121,160,:,:,0].shape)
	#_nrows = int(ijt.image.shape[0] / ijt.tileH)
	#_ncols = int(ijt.image.shape[1] / ijt.tileW)
	#_nrows = 32
	#_ncols = 32
	#print(str(_nrows)+" "+str(_ncols))
	#fig, ax = plt.subplots(nrows=_nrows, ncols=_ncols)
	#for i in range(_nrows):
	# for j in range(_ncols):
	# ax[i,j].imshow(tiles[i+100,j,:,:,0])
	# ax[i,j].set_axis_off()

	#for i in range(5):
	# fig = plt.figure()
	# plt.imshow(tiles[121,160,:,:,i])
	# plt.colorbar()

	#ijt.show_images(['other'])

	#ijt.show_images([0,3])
	#ijt.show_images(['X-corr','Y-corr'])
	#ijt.show_images(['R-vign',3])

	ijt.show_images()		ijt.show_images()
	plt.show()		plt.show()

Admin message