def export(mesh_file, agg_file, seg_file, label_map_file, type, output_file):
label_map = util.read_label_mapping(opt.label_map_file,
label_from='raw_category',
label_to='nyu40id')
mesh_vertices = util_3d.read_mesh_vertices(mesh_file)
object_id_to_segs, label_to_segs = read_aggregation(agg_file)
seg_to_verts, num_verts = read_segmentation(seg_file)
label_ids = np.zeros(shape=(num_verts), dtype=np.uint32) # 0: unannotated
for label, segs in label_to_segs.items():
label_id = label_map[label]
for seg in segs:
verts = seg_to_verts[seg]
label_ids[verts] = label_id
if type == 'label':
util_3d.export_ids(output_file, label_ids)
elif type == 'instance':
instance_ids = np.zeros(shape=(num_verts),
dtype=np.uint32) # 0: unannotated
for object_id, segs in object_id_to_segs.items():
for seg in segs:
verts = seg_to_verts[seg]
instance_ids[verts] = object_id
util_3d.export_ids(output_file, label_ids * 1000 + instance_ids)
#util_3d.export_instance_ids_for_eval(output_file, label_ids, instance_ids)
else:
raise
def export(mesh_file, agg_file, seg_file, label_map_file, label_map_file2, type, output_file):
label_map = util.read_label_mapping(opt.label_map_file, label_from='raw_category', label_to='nyu40id')
label_map2 = load_mapping(label_map_file2)
label_map2[0] = 0
mesh_vertices = util_3d.read_mesh_vertices(mesh_file)
object_id_to_segs, label_to_segs = read_aggregation(agg_file)
seg_to_verts, num_verts = read_segmentation(seg_file)
label_ids = np.zeros(shape=(num_verts), dtype=np.uint32) # 0: unannotated
for label, segs in label_to_segs.items():
label_id = label_map[label]
for seg in segs:
verts = seg_to_verts[seg]
label_ids[verts] = label_id
if type == 'label':
util_3d.export_ids(output_file, label_ids)
elif type == 'instance':
instance_ids = np.zeros(shape=(num_verts), dtype=np.uint32) # 0: unannotated
for object_id, segs in object_id_to_segs.items():
for seg in segs:
verts = seg_to_verts[seg]
instance_ids[verts] = object_id
assert len(instance_ids) == len(label_ids)
for i in range(len(instance_ids)):
import ipdb
ipdb.set_trace()
instance_ids[i] = label_map2[label_ids[i]] * 1000 + instance_ids[i]
util_3d.export_ids(output_file, instance_ids)
else:
raise
def main():
image = np.array(imageio.imread(opt.input_file))
label_map = util.read_label_mapping(opt.label_map_file,
label_from='id',
label_to='nyu40id')
mapped_image = map_label_image(image, label_map)
imageio.imwrite(opt.output_file, mapped_image)
def main():
instance_image = np.array(imageio.imread(opt.input_instance_file))
label_image = np.array(imageio.imread(opt.input_label_file))
label_map = util.read_label_mapping(opt.label_map_file,
label_from='id',
label_to='nyu40id')
mapped_label = map_label_image(label_image, label_map)
output_instance_image = make_instance_image(mapped_label, instance_image)
imageio.imwrite(opt.output_file, output_instance_image)
def main():
if not os.path.exists(opt.output_path):
os.makedirs(opt.output_path)
label_mapping = None
if opt.export_label_images:
label_map = util.read_label_mapping(opt.label_map_file, label_from='id', label_to='nyu40id')
scenes = [d for d in os.listdir(opt.scannet_path) if os.path.isdir(os.path.join(opt.scannet_path, d))]
print('Found %d scenes' % len(scenes))
for i in range(len(scenes)):
sens_file = os.path.join(opt.scannet_path, scenes[i], scenes[i] + '.sens')
label_path = os.path.join(opt.scannet_path, scenes[i], opt.label_type)
if opt.export_label_images and not os.path.isdir(label_path):
print_error('Error: using export_label_images option but label path %s does not exist' % label_path)
output_color_path = os.path.join(opt.output_path, scenes[i], 'color')
if not os.path.isdir(output_color_path):
os.makedirs(output_color_path)
output_depth_path = os.path.join(opt.output_path, scenes[i], 'depth')
if not os.path.isdir(output_depth_path):
os.makedirs(output_depth_path)
output_pose_path = os.path.join(opt.output_path, scenes[i], 'pose')
if not os.path.isdir(output_pose_path):
os.makedirs(output_pose_path)
output_label_path = os.path.join(opt.output_path, scenes[i], 'label')
if opt.export_label_images and not os.path.isdir(output_label_path):
os.makedirs(output_label_path)
# read and export
sys.stdout.write('\r[ %d | %d ] %s\tloading...' % ((i + 1), len(scenes), scenes[i]))
sys.stdout.flush()
sd = SensorData(sens_file)
sys.stdout.write('\r[ %d | %d ] %s\texporting...' % ((i + 1), len(scenes), scenes[i]))
sys.stdout.flush()
sd.export_color_images(output_color_path, image_size=[opt.output_image_height, opt.output_image_width], frame_skip=opt.frame_skip)
sd.export_depth_images(output_depth_path, image_size=[opt.output_image_height, opt.output_image_width], frame_skip=opt.frame_skip)
sd.export_poses(output_pose_path, frame_skip=opt.frame_skip)
if opt.export_label_images:
for f in range(0, len(sd.frames), opt.frame_skip):
label_file = os.path.join(label_path, str(f) + '.png')
image = np.array(imageio.imread(label_file))
image = sktf.resize(image, [opt.output_image_height, opt.output_image_width], order=0, preserve_range=True)
mapped_image = map_label_image(image, label_map)
imageio.imwrite(os.path.join(output_label_path, str(f) + '.png'), mapped_image)
print('')
def main():
if not os.path.exists(opt.output_path):
os.makedirs(opt.output_path)
label_mapping = None
if opt.export_label_images:
label_map = util.read_label_mapping(opt.label_map_file, label_from='id', label_to='nyu40id')
scenes = [d for d in os.listdir(opt.scannet_path) if os.path.isdir(os.path.join(opt.scannet_path, d))]
print('Found %d scenes' % len(scenes))
print(scenes[:5])
# for i in range(len(scenes)):
combined_x = [(scenes[i], i, len(scenes), label_map) for i in range(len(scenes))]
with Pool(processes=32, initializer=randomize) as pool:
# _ = list(tqdm(pool.imap_unordered(process, zip(scenes, range(len(scenes)), [len(scenes)]*len(scenes))), total=len(scenes)))
_ = list(tqdm(pool.imap_unordered(process, combined_x), total=len(scenes)))
print('')
'--scan_path',
required=True,
help='path to scannet scene (e.g., data/ScanNet/v2/scene0000_00')
parser.add_argument('--output_file', required=True, help='output file')
parser.add_argument(
'--label_map_file',
default='/gruvi/Data/chenliu/ScanNet/tasks/scannetv2-labels.combined.tsv',
help='path to scannetv2-labels.combined.tsv')
parser.add_argument('--type',
default='instance',
help='task type [label or instance]')
opt = parser.parse_args()
assert opt.type in TASK_TYPES
label_map = util.read_label_mapping(opt.label_map_file,
label_from='raw_category',
label_to='nyu40id')
# remapper=np.ones(150)*(-100)
# for i,x in enumerate([1,2,3,4,5,6,7,8,9,10,11,12,14,16,24,28,33,34,36,39]):
# remapper[x]=i
def read_aggregation(filename):
assert os.path.isfile(filename)
object_id_to_segs = {}
label_to_segs = {}
with open(filename) as f:
data = json.load(f)
num_objects = len(data['segGroups'])
for i in range(num_objects):
object_id = data['segGroups'][i][
def main():
if not os.path.exists(opt.output_path):
os.makedirs(opt.output_path)
label_mapping = None
if opt.export_label_images:
label_map = util.read_label_mapping(opt.label_map_file, label_from='id', label_to='nyu40id')
scenes = [d for d in os.listdir(opt.scannet_path) if os.path.isdir(os.path.join(opt.scannet_path, d))]
print('Found %d scenes' % len(scenes))
for i in range(len(scenes)):
sens_file = os.path.join(opt.scannet_path, scenes[i], scenes[i] + '.sens')
label_path = os.path.join(opt.scannet_path, scenes[i], opt.label_type)
if opt.export_label_images and not os.path.isdir(label_path):
print_error('Error: using export_label_images option but label path %s does not exist' % label_path)
# Variable to indicate what kind of sensor data is to be extracted
reextract = []
output_color_path = os.path.join(opt.output_path, scenes[i], 'color')
if not os.path.isdir(output_color_path):
os.makedirs(output_color_path)
reextract.append('color')
output_depth_path = os.path.join(opt.output_path, scenes[i], 'depth')
if not os.path.isdir(output_depth_path):
os.makedirs(output_depth_path)
reextract.append('depth')
output_pose_path = os.path.join(opt.output_path, scenes[i], 'pose')
if not os.path.isdir(output_pose_path):
os.makedirs(output_pose_path)
reextract.append('pose')
output_label_path = os.path.join(opt.output_path, scenes[i], 'label')
if opt.export_label_images and not os.path.isdir(output_label_path):
os.makedirs(output_label_path)
reextract.append('label')
# If we only have to extract 'label', we can skip exporting sensor data
# We can also continue to the next scene if we do not have to reextract anything
if (len(reextract) == 1 and reextract[0] == 'label') or reextract == []:
sys.stdout.write('\r[ %d | %d ] %s\tskipping sensordata...' % ((i+1), len(scenes), scenes[i]))
sys.stdout.flush()
if reextract == []:
continue
# read and export
sys.stdout.write('\r[ %d | %d ] %s\tloading...' % ((i + 1), len(scenes), scenes[i]))
sys.stdout.flush()
print(sens_file)
sd = SensorData(sens_file)
else:
sys.stdout.write('\r[ %d | %d ] %s\texporting...' % ((i + 1), len(scenes), scenes[i]))
sys.stdout.flush()
sd.export_color_images(output_color_path, image_size=[opt.output_image_height, opt.output_image_width], frame_skip=opt.frame_skip)
sd.export_depth_images(output_depth_path, image_size=[opt.output_image_height, opt.output_image_width], frame_skip=opt.frame_skip)
sd.export_poses(output_pose_path, frame_skip=opt.frame_skip)
if opt.export_label_images:
for f in range(0, len(sd.frames), opt.frame_skip):
label_file = os.path.join(label_path, str(f) + '.png')
image = np.array(imageio.imread(label_file))
image = sktf.resize(image, [opt.output_image_height, opt.output_image_width], order=0, preserve_range=True)
mapped_image = map_label_image(image, label_map)
imageio.imwrite(os.path.join(output_label_path, str(f) + '.png'), mapped_image)
print('')