def __init__(self):
AlgorithmPluginManager.register_plugins()
# Algorithms
self.algo_convert_img = ConvertImage("convert_image")
self.algo_image_io = ImageIo("image_reader")
self.algo_track_features = TrackFeatures("feature_tracker")
# Other tool variables
self.image_list_filepath = None
self.mask_list_filepath = None
self.invert_masks = False
self.expect_multichannel_masks = False
self.output_tracks_filepath = None
def __init__(self):
AlgorithmPluginManager.register_plugins()
# Algorithms
self.algo_convert_img = ConvertImage("convert_image")
self.algo_image_io = ImageIo("image_reader")
self.algo_track_features = TrackFeatures("feature_tracker")
# Other tool variables
self.image_list_filepath = None
self.mask_list_filepath = None
self.invert_masks = False
self.expect_multichannel_masks = False
self.output_tracks_filepath = None
def test_image_load_save_diff(self):
fd, tmp_filename = tempfile.mkstemp()
c = ConfigBlock()
c.set_value('iio:type', 'vxl')
iio = ImageIo('iio')
iio.set_config(c)
nt.assert_true(osp.isfile(self.test_image_filepath),
"Couldn't find image file")
ic_orig = iio.load(self.test_image_filepath)
iio.save(ic_orig, tmp_filename)
ic_test = iio.load(tmp_filename)
nt.assert_equal(ic_orig.size(), ic_test.size())
nt.assert_equal(ic_orig.width(), ic_test.width())
nt.assert_equal(ic_orig.height(), ic_test.height())
nt.assert_equal(ic_orig.depth(), ic_test.depth())
os.remove(tmp_filename)
os.close(fd)
def test_image_load_save_diff(self):
fd, tmp_filename = tempfile.mkstemp()
c = ConfigBlock()
c.set_value("iio:type", "vxl")
iio = ImageIo("iio")
iio.set_config(c)
nt.assert_true(osp.isfile(self.test_image_filepath), "Couldn't find image file")
ic_orig = iio.load(self.test_image_filepath)
iio.save(ic_orig, tmp_filename)
ic_test = iio.load(tmp_filename)
nt.assert_equal(ic_orig.size(), ic_test.size())
nt.assert_equal(ic_orig.width(), ic_test.width())
nt.assert_equal(ic_orig.height(), ic_test.height())
nt.assert_equal(ic_orig.depth(), ic_test.depth())
os.remove(tmp_filename)
os.close(fd)
class TrackFeaturesTool (object):
@property
def log(self):
return logging.getLogger("TrackFeaturesTool")
def __init__(self):
AlgorithmPluginManager.register_plugins()
# Algorithms
self.algo_convert_img = ConvertImage("convert_image")
self.algo_image_io = ImageIo("image_reader")
self.algo_track_features = TrackFeatures("feature_tracker")
# Other tool variables
self.image_list_filepath = None
self.mask_list_filepath = None
self.invert_masks = False
self.expect_multichannel_masks = False
self.output_tracks_filepath = None
def get_configuration(self):
"""
:return: the current tool configuration
:rtype: ConfigBlock
"""
c = ConfigBlock("track_features")
c.set_value("image_list_file",
self.image_list_filepath or "",
"Path to an input file containing new-line separated paths "
"to sequential image files")
c.set_value("mask_list_file",
self.mask_list_filepath or "",
"Optional path to an input file containing new-line "
"separated paths to mask images. This list should be "
"parallel in association to files specified in "
"``image_list_file``. Mask image must be the same size as "
"the image they are associated with.\n"
"\n"
"Leave this blank if no image masking is desired.")
c.set_value("invert_masks",
str(self.invert_masks).lower(),
"If true, all mask images will be inverted after loading. "
"This is useful if mask images read in use positive "
"values to indicated masked areas instead of non-masked "
"areas.")
c.set_value("expect_multichannel_masks",
str(self.expect_multichannel_masks).lower(),
"A majority of the time, mask images are a single channel, "
"however it is feasibly possible that certain "
"implementations may use multi-channel masks. If this is "
"true we will expect multiple-channel mask images, "
"warning when a single-channel mask is provided. If this "
"is false we error upon seeing a multi-channel mask "
"image.")
c.set_value("output_tracks_file",
self.output_tracks_filepath or "",
"Path to a file to write output tracks to. If this "
"file exists, it will be overwritten.")
# Required algorithm does not have an implemented interface yet
# c.set_value("output_homography_file",
# self.output_homography_filepath or "",
# "Optional path to a file to write source-to-reference "
# "homographies for each frame. Leave blank to disable this "
# "output. The output_homography_generator algorithm type "
# "only needs to be set if this is set.")
self.algo_convert_img.get_config(c)
self.algo_image_io.get_config(c)
self.algo_track_features.get_config(c)
return c
def set_configuration(self, config):
"""
Set the tool configuration
:param config: The configuration to set
:type config: ConfigBlock
"""
self.algo_convert_img.set_config(config)
self.algo_image_io.set_config(config)
self.algo_track_features.set_config(config)
abspath = lambda p: os.path.abspath(os.path.expanduser(p))
self.image_list_filepath = config.get_value('image_list_file',
self.image_list_filepath)
if self.image_list_filepath:
self.image_list_filepath = abspath(self.image_list_filepath)
self.mask_list_filepath = config.get_value('mask_list_file',
self.mask_list_filepath)
if self.mask_list_filepath:
self.mask_list_filepath = abspath(self.mask_list_filepath)
self.invert_masks = config.get_value_bool('invert_masks', self.invert_masks)
self.expect_multichannel_masks = \
config.get_value_bool('expect_multichannel_masks',
self.expect_multichannel_masks)
self.output_tracks_filepath = \
config.get_value('output_tracks_file', self.output_tracks_filepath)
if self.output_tracks_filepath:
self.output_tracks_filepath = abspath(self.output_tracks_filepath)
def validate_configuration(self):
"""
Check the current configuration for validity
:return: True if valid, false if not.
:type: bool
"""
if not self.image_list_filepath:
self.log.error("No image list file provided")
return False
if not os.path.isfile(self.image_list_filepath):
self.log.error("Path given for image list does not refer to a file")
return False
if not self.output_tracks_filepath:
self.log.error("No output tracks file privided")
return False
if self.mask_list_filepath and not os.path.isfile(self.mask_list_filepath):
self.log.error("Specified a mask list filepath, but path did not "
"refer to an existing file.")
return False
return (
# Algorithms should check out valid
self.algo_convert_img.check_config(self.algo_convert_img.get_config())
and self.algo_image_io.check_config(self.algo_image_io.get_config())
and self.algo_track_features.check_config(self.algo_track_features.get_config())
)
def track_images(self):
# Create list of input filepaths
self.log.info("Reading input image list file...")
with open(self.image_list_filepath, 'r') as image_list_file:
input_filepaths = \
[line.strip() for line in image_list_file.readlines()]
# Create a list of mask images if we were given a list file
mask_filepaths = None
if self.mask_list_filepath:
self.log.info("Reading mask image list file...")
with open(self.mask_list_filepath) as mask_list_file:
mask_filepaths = \
[line.strip() for line in mask_list_file.readlines()]
if len(input_filepaths) != len(mask_filepaths):
self.log.error("Input and Mask image lists were not congruent "
"in size.")
return False
# Check that the output tracks file is open-able and that the containing
# directory exists.
if not os.path.isdir(os.path.dirname(self.output_tracks_filepath)):
self.log.info("Creating containing directory for output tracks "
"file: %s",
os.path.dirname(self.output_tracks_filepath))
os.mkdir(os.path.dirname(self.output_tracks_filepath))
self.log.info("Testing that output tracks file is open-able...")
test_f = open(self.output_tracks_filepath, 'w')
test_f.close()
tracks = TrackSet()
for frame_num in xrange(len(input_filepaths)):
input_img = self.algo_convert_img.convert(
self.algo_image_io.load(input_filepaths[frame_num])
)
mask_img = None
if mask_filepaths:
mask_img = self.algo_convert_img.convert(
self.algo_image_io.load(mask_filepaths[frame_num])
)
self.log.info("Processing frame %d...", frame_num)
tracks = self.algo_track_features.track(tracks, frame_num,
input_img, mask_img)
self.log.info("Frame processing complete, writing out track set...")
tracks.write_tracks_file(self.output_tracks_filepath)
class TrackFeaturesTool(object):
@property
def log(self):
return logging.getLogger("TrackFeaturesTool")
def __init__(self):
AlgorithmPluginManager.register_plugins()
# Algorithms
self.algo_convert_img = ConvertImage("convert_image")
self.algo_image_io = ImageIo("image_reader")
self.algo_track_features = TrackFeatures("feature_tracker")
# Other tool variables
self.image_list_filepath = None
self.mask_list_filepath = None
self.invert_masks = False
self.expect_multichannel_masks = False
self.output_tracks_filepath = None
def get_configuration(self):
"""
:return: the current tool configuration
:rtype: ConfigBlock
"""
c = ConfigBlock("track_features")
c.set_value(
"image_list_file", self.image_list_filepath or "",
"Path to an input file containing new-line separated paths "
"to sequential image files")
c.set_value(
"mask_list_file", self.mask_list_filepath or "",
"Optional path to an input file containing new-line "
"separated paths to mask images. This list should be "
"parallel in association to files specified in "
"``image_list_file``. Mask image must be the same size as "
"the image they are associated with.\n"
"\n"
"Leave this blank if no image masking is desired.")
c.set_value(
"invert_masks",
str(self.invert_masks).lower(),
"If true, all mask images will be inverted after loading. "
"This is useful if mask images read in use positive "
"values to indicated masked areas instead of non-masked "
"areas.")
c.set_value(
"expect_multichannel_masks",
str(self.expect_multichannel_masks).lower(),
"A majority of the time, mask images are a single channel, "
"however it is feasibly possible that certain "
"implementations may use multi-channel masks. If this is "
"true we will expect multiple-channel mask images, "
"warning when a single-channel mask is provided. If this "
"is false we error upon seeing a multi-channel mask "
"image.")
c.set_value(
"output_tracks_file", self.output_tracks_filepath or "",
"Path to a file to write output tracks to. If this "
"file exists, it will be overwritten.")
# Required algorithm does not have an implemented interface yet
# c.set_value("output_homography_file",
# self.output_homography_filepath or "",
# "Optional path to a file to write source-to-reference "
# "homographies for each frame. Leave blank to disable this "
# "output. The output_homography_generator algorithm type "
# "only needs to be set if this is set.")
self.algo_convert_img.get_config(c)
self.algo_image_io.get_config(c)
self.algo_track_features.get_config(c)
return c
def set_configuration(self, config):
"""
Set the tool configuration
:param config: The configuration to set
:type config: ConfigBlock
"""
self.algo_convert_img.set_config(config)
self.algo_image_io.set_config(config)
self.algo_track_features.set_config(config)
abspath = lambda p: os.path.abspath(os.path.expanduser(p))
self.image_list_filepath = config.get_value('image_list_file',
self.image_list_filepath)
if self.image_list_filepath:
self.image_list_filepath = abspath(self.image_list_filepath)
self.mask_list_filepath = config.get_value('mask_list_file',
self.mask_list_filepath)
if self.mask_list_filepath:
self.mask_list_filepath = abspath(self.mask_list_filepath)
self.invert_masks = config.get_value_bool('invert_masks',
self.invert_masks)
self.expect_multichannel_masks = \
config.get_value_bool('expect_multichannel_masks',
self.expect_multichannel_masks)
self.output_tracks_filepath = \
config.get_value('output_tracks_file', self.output_tracks_filepath)
if self.output_tracks_filepath:
self.output_tracks_filepath = abspath(self.output_tracks_filepath)
def validate_configuration(self):
"""
Check the current configuration for validity
:return: True if valid, false if not.
:type: bool
"""
if not self.image_list_filepath:
self.log.error("No image list file provided")
return False
if not os.path.isfile(self.image_list_filepath):
self.log.error(
"Path given for image list does not refer to a file")
return False
if not self.output_tracks_filepath:
self.log.error("No output tracks file privided")
return False
if self.mask_list_filepath and not os.path.isfile(
self.mask_list_filepath):
self.log.error("Specified a mask list filepath, but path did not "
"refer to an existing file.")
return False
return (
# Algorithms should check out valid
self.algo_convert_img.check_config(
self.algo_convert_img.get_config()) and
self.algo_image_io.check_config(self.algo_image_io.get_config())
and self.algo_track_features.check_config(
self.algo_track_features.get_config()))
def track_images(self):
# Create list of input filepaths
self.log.info("Reading input image list file...")
with open(self.image_list_filepath, 'r') as image_list_file:
input_filepaths = \
[line.strip() for line in image_list_file.readlines()]
# Create a list of mask images if we were given a list file
mask_filepaths = None
if self.mask_list_filepath:
self.log.info("Reading mask image list file...")
with open(self.mask_list_filepath) as mask_list_file:
mask_filepaths = \
[line.strip() for line in mask_list_file.readlines()]
if len(input_filepaths) != len(mask_filepaths):
self.log.error("Input and Mask image lists were not congruent "
"in size.")
return False
# Check that the output tracks file is open-able and that the containing
# directory exists.
if not os.path.isdir(os.path.dirname(self.output_tracks_filepath)):
self.log.info(
"Creating containing directory for output tracks "
"file: %s", os.path.dirname(self.output_tracks_filepath))
os.mkdir(os.path.dirname(self.output_tracks_filepath))
self.log.info("Testing that output tracks file is open-able...")
test_f = open(self.output_tracks_filepath, 'w')
test_f.close()
tracks = TrackSet()
for frame_num in xrange(len(input_filepaths)):
input_img = self.algo_convert_img.convert(
self.algo_image_io.load(input_filepaths[frame_num]))
mask_img = None
if mask_filepaths:
mask_img = self.algo_convert_img.convert(
self.algo_image_io.load(mask_filepaths[frame_num]))
self.log.info("Processing frame %d...", frame_num)
tracks = self.algo_track_features.track(tracks, frame_num,
input_img, mask_img)
self.log.info("Frame processing complete, writing out track set...")
tracks.write_tracks_file(self.output_tracks_filepath)
