def load_specific_dataset(self):
dataset_config_filename = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
'dataset', 'composite', 'hats_3_demo_composite.yaml')
dataset_config_filename = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence',
'dataset', 'composite', '4_shoes_all.yaml')
dataset_config = utils.getDictFromYamlFilename(dataset_config_filename)
self._dataset = SpartanDataset(config=dataset_config)
def load_specific_dataset(self):
dataset_config_filename = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'dataset', 'composite',
'hats_3_demo_composite.yaml')
dataset_config_filename = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'dataset', 'composite', '4_shoes_all.yaml')
dataset_config = utils.getDictFromYamlFilename(dataset_config_filename)
self._dataset = SpartanDataset(config=dataset_config)
def load_default_config():
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
config_file = os.path.join(dc_source_dir, 'config', 'dense_correspondence',
'training', 'training.yaml')
config = utils.getDictFromYamlFilename(config_file)
return config
def load_default_config():
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
config_file = os.path.join(dc_source_dir, 'config', 'dense_correspondence',
'training', 'training.yaml')
config = utils.getDictFromYamlFilename(config_file)
return config
def load_specific_dataset(self):
dataset_config_filename = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'dataset', 'composite',
'rope_nonrigid_412vert_only.yaml')
dataset_config = utils.getDictFromYamlFilename(dataset_config_filename)
self._dataset = SpartanDataset(config=dataset_config)
def get_descriptor_target_from_yaml(self):
"""
Grabs a 1-dimensional numpy array of length D from the descriptor yaml file
"""
descriptor_filename = os.path.join(pdc_utils.getDenseCorrespondenceSourceDir(), "../config", "new_descriptor_picked.yaml")
descriptor_dict = pdc_utils.getDictFromYamlFilename(descriptor_filename)
descriptor_list = descriptor_dict["descriptor"]
return np.asarray(descriptor_list)
def get_unet(config):
"""
Returns a Unet nn.module that satisifies the fcn properties stated in get_fcn() docstring
"""
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
sys.path.append(os.path.join(dc_source_dir, 'external/unet-pytorch'))
from unet_model import UNet
model = UNet(num_classes=config["descriptor_dimension"]).cuda()
return model
def make_default_caterpillar():
"""
Makes a default SpartanDatase from the 10_scenes_drill data
:return:
:rtype:
"""
config_file = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
'dataset', 'composite',
'caterpillar_only.yaml')
config = utils.getDictFromYamlFilename(config_file)
dataset = SpartanDataset(mode="train", config=config)
return dataset
def __init__(self, config_filename='shoes_all.yaml'):
with HiddenPrints():
self.config_filename = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'dataset', 'composite',
config_filename)
self.train_config_filename = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'training', 'training.yaml')
self.config = utils.getDictFromYamlFilename(self.config_filename)
self.train_config = utils.getDictFromYamlFilename(
self.train_config_filename)
self.dataset = SpartanDataset(config=self.config)
self.dataset.set_parameters_from_training_config(self.train_config)
# holds centroid and radius for each scene
# these are for min and max z values currently. maybe include x, y, and z in the future.
# self.centroid_and_radius[scene_name]["centroid"] or self.centroid_and_radius[scene_name]["radius"]
self.centroid_and_radius = {}
def run(data_folder, ply_binary_filename='images.ply'):
# install ply if do not already have it
os.chdir(data_folder)
vtp_filename = os.path.join(data_folder, 'images.vtp')
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
ply_to_ascii_executable = os.path.join(dc_source_dir, 'src', 'ply',
'ply2ascii')
path_to_ply = os.path.join(dc_source_dir, "src", "ply")
if not (os.path.isfile(ply_to_ascii_executable)):
os.system("cd " + path_to_ply + " && make")
correct_ply_header_file = os.path.join(dc_source_dir, 'config',
'correct_ply_header.txt')
ply_binary_full_filename = os.path.join(data_folder, ply_binary_filename)
converted_ascii_filename = os.path.join(data_folder,
"converted_to_ascii.ply")
converted_ascii_modified_header_filename = os.path.join(
data_folder, "converted_to_ascii_modified_header.ply")
# call ply2ascii
os.system(ply_to_ascii_executable + "<./" + ply_binary_filename + "> " +
converted_ascii_filename)
# change header to be compatible with Director
# TODO: make so Director accepts other header?
line_elements_vertex = ""
with open(converted_ascii_modified_header_filename, 'w') as outfile:
with open(converted_ascii_filename) as infile:
counter = 0
for line in infile:
counter += 1
if counter == 3:
line_elements_vertex = line
break
with open(correct_ply_header_file) as infile:
counter = 0
for line in infile:
counter += 1
if counter == 4:
outfile.write(line_elements_vertex)
continue
outfile.write(line)
with open(converted_ascii_filename) as infile:
num_skip = 14
counter = 0
for line in infile:
counter += 1
if counter <= 14:
continue
outfile.write(line)
# convert to vtp
convert_ply_to_vtp_script = os.path.join(
dc_source_dir, 'modules', 'dense_correspondence_manipulation',
'scripts', 'convertPlyToVtp.py')
print "converted to ascii ply format"
os.system("directorPython " + convert_ply_to_vtp_script + " " +
converted_ascii_modified_header_filename)
converted_ascii_modified_header_vtp_filename = os.path.join(
data_folder, "converted_to_ascii_modified_header.vtp")
print "finished convert_ply_to_vtp_script"
# clean up and rename
# os.system("rm *.ply *.freiburg")
os.rename(converted_ascii_modified_header_vtp_filename, vtp_filename)
import numpy as np
# pdc
import dense_correspondence_manipulation.utils.utils as utils
from dense_correspondence.evaluation.evaluation import DenseCorrespondenceEvaluation
from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset
from dense_correspondence.dataset.scene_structure import SceneStructure
"""
Computes descriptor images for a given scene and network. Saves them as
npy files.
Usage: Modify the global variables in CAPS as needed
"""
DC_SOURCE_DIR = utils.getDenseCorrespondenceSourceDir()
NETWORK_NAME = "caterpillar_M_background_0.500_3"
EVALUATION_CONFIG_FILENAME = os.path.join(DC_SOURCE_DIR, 'config',
'dense_correspondence', 'evaluation',
'lucas_evaluation.yaml')
DATASET_CONFIG_FILE = os.path.join(DC_SOURCE_DIR, 'config',
'dense_correspondence', 'dataset',
'composite', 'caterpillar_only_9.yaml')
SCENE_NAME = "2018-04-16-14-25-19"
SAVE_DIR = os.path.join("/home/manuelli/code/data_volume/pdc/logs_test",
SCENE_NAME, "processed", "descriptor_images",
NETWORK_NAME)
import os
import dense_correspondence_manipulation.utils.utils as utils
CHANGE_DETECTION_CONFIG_FILE = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'stations', 'RLG_iiwa_1', 'change_detection.yaml')
CHANGE_DETECTION_BACKGROUND_SUBTRACTION_CONFIG_FILE = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'stations', 'RLG_iiwa_1', 'change_detection_background_subtraction.yaml')
BACKGROUND_SCENE_DATA_FOLDER = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'data_volume', 'pdc','logs_proto', '14_background')
DEPTH_IM_SCALE = 1000.0 # This represents that depth images are saved as uint16, where the integer value
# is depth in millimeters. So this scale just converts millimeters to meters.
DEPTH_IM_RESCALE = 4000.0 # Only for visualizaiton purposes
IMAGE_NET_MEAN = [0.485, 0.456, 0.406]
IMAGE_NET_STD_DEV = [0.229, 0.224, 0.225]
DEFAULT_IMAGE_MEAN = [0.5573105812072754, 0.37420374155044556, 0.37020164728164673]
DEFAULT_IMAGE_STD_DEV = [0.24336038529872894, 0.2987397611141205, 0.31875079870224]
if not debug:
TimerCallback(callback=single_shot_function).singleShot(0)
app.app.start(restoreWindow=True)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument(
"--data_dir",
type=str,
default=
'/home/manuelli/code/data_volume/sandbox/drill_scenes/04_drill_long_downsampled'
)
default_config_file = os.path.join(utils.getDenseCorrespondenceSourceDir(),
'config', 'stations', 'RLG_iiwa_1',
'change_detection.yaml')
parser.add_argument("--config_file", type=str, default=default_config_file)
parser.add_argument(
'--current_dir',
action='store_true',
default=False,
help="run the script with --data_dir set to the current directory")
parser.add_argument('--debug',
action='store_true',
default=False,
help="launch the app in debug mode")
from utils import *
# pdc_ros_msgs
import pdc_ros_msgs.msg
import time
import sys
import cv2
from std_msgs.msg import String
from sensor_msgs.msg import Image
from cv_bridge import CvBridge, CvBridgeError
from PIL import Image as PILImage
import dense_correspondence_manipulation.utils.utils as pdc_utils
pdc_utils.add_dense_correspondence_to_python_path()
dc_source_dir = pdc_utils.getDenseCorrespondenceSourceDir()
from dense_correspondence.evaluation.evaluation import DenseCorrespondenceEvaluation
from dense_correspondence.evaluation.plotting import normalize_descriptor
import torch
NETWORK_CONFIG_FILENAME = os.path.join(get_config_directory(),
'trained_networks.yaml')
NETWORK_NAME = "shoes_consistent_M_background_0.500_3"
RGB_TOPIC = "/camera_carmine_1/rgb/image_rect_color"
class StreamingPdcRos(object):
def __init__(self):
self.bridge = None
self.load_dcn_network()
import sys import os import cv2 import numpy as np import copy import dense_correspondence_manipulation.utils.utils as utils dc_source_dir = utils.getDenseCorrespondenceSourceDir() sys.path.append(dc_source_dir) sys.path.append(os.path.join(dc_source_dir, "dense_correspondence", "correspondence_tools")) from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset, ImageType import dense_correspondence from dense_correspondence.evaluation.evaluation import * from dense_correspondence.evaluation.plotting import normalize_descriptor from dense_correspondence.network.dense_correspondence_network import DenseCorrespondenceNetwork sys.path.append(os.path.join(os.path.dirname(__file__), "../simple-pixel-correspondence-labeler")) from annotate_correspondences import label_colors, draw_reticle, pil_image_to_cv2, drawing_scale_config, numpy_to_cv2, label_colors COLOR_RED = np.array([0, 0, 255]) COLOR_GREEN = np.array([0,255,0]) utils.set_default_cuda_visible_devices() eval_config_filename = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence', 'evaluation', 'lucas_evaluation.yaml') EVAL_CONFIG = utils.getDictFromYamlFilename(eval_config_filename) class HeatmapVisualization(object):
import sys
import os
import cv2
import numpy as np
import copy
import dense_correspondence_manipulation.utils.utils as utils
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
sys.path.append(dc_source_dir)
sys.path.append(os.path.join(dc_source_dir, "dense_correspondence", "correspondence_tools"))
from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset, ImageType
config_filename = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
'dataset', 'composite', 'star_bot_front_only.yaml')
config = utils.getDictFromYamlFilename(config_filename)
sd = SpartanDataset(config=config)
sd.set_train_mode()
USE_FIRST_IMAGE = False # force using first image in each log
RANDOMIZE_TEST_TRAIN = False # randomize seletcting
def numpy_to_cv2(numpy_img):
return numpy_img[:, :, ::-1].copy() # open and convert between BGR and RGB
def pil_image_to_cv2(pil_image):
return np.array(pil_image)[:, :, ::-1].copy() # open and convert between BGR and RGB
def get_cv2_img_pair_from_spartan():
scene_name_a = sd.get_random_scene_name()
num_attempts = 50
for i in range(num_attempts):
if not os.path.isdir(processed_dir):
raise ValueError("Need to extract and run fusion on " + full_dir)
if already_ran_change_detection(processed_dir):
print "already_ran_change_detection for", processed_dir
continue
# print "full_dir", full_dir
#print "would have run on ", full_dir
run(processed_dir, config_file)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--data_dir", type=str, default='/home/manuelli/code/data_volume/sandbox/drill_scenes/01_drill')
default_config_file = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'stations', 'RLG_iiwa_1', 'change_detection.yaml')
parser.add_argument("--config_file", type=str, default=default_config_file)
parser.add_argument('--current_dir', action='store_true', default=True, help="run the script with --data_dir set to the current directory")
args = parser.parse_args()
data_folder = args.data_dir
if args.current_dir:
print "running with data_dir set to current working directory . . . "
data_folder = os.getcwd()
# run(data_folder, config_file=args.config_file)
run_on_all_subfolders(data_folder, args.config_file)
def __init__(self, config_path):
# manuly write configs, debug use only
self.config = {}
self.config['generate_dataset'] = {}
self.config['generate_dataset']['required'] = False
self.config['generate_dataset']['generate_depth'] = True
self.config['generate_dataset']['data_source'] = '../data/pdc'
self.config['generate_dataset'][
'original_data'] = 'logs_proto_original'
self.config['generate_dataset']['meta_dir'] = None
self.config['generate_dataset']['output_target'] = 'rendered_images'
self.config['generate_dataset']['image_dir'] = None
self.config['generate_dataset']['output_dir'] = None
self.config['generate_dataset'][
'model_path'] = '../data/pdc/depth_models/weights_199'
self.config['generate_dataset']['scaling_method'] = 'default_scaling'
self.config['generate_dataset']['zero_masked'] = False
self.config['generate_dataset']['ext'] = 'png'
self.config['generate_dataset']['no_cuda'] = False
self.config['generate_dataset']['replace_poses'] = False
self.config['generate_dataset']['pose_data_path'] = '../data/pdc/poses'
self.config['train'] = {}
self.config['train']['required'] = True
# self.config['train']['dataset_config_file'] = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
# 'dataset', 'composite', 'caterpillar_upright.yaml')
self.config['train']['dataset_config_file'] = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config', 'kitti',
'kitti.yaml')
self.config['train']['train_config_file'] = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'training', 'training.yaml')
self.config['train'][
'logging_dir'] = "trained_models/new_new_test_caterpillar/"
self.config['train']['num_iterations'] = (1500 / 4) - 1
self.config['train']['dimension'] = 3
# self.config['train']['dataset'] = "logs_proto_original"
# self.config['train']['dataset'] = "experiments/exp_07232020-184440/logs_proto_default_scaling_gt_pose"
# self.config['train']['dataset'] = "experiments/exp_05262020-174446/logs_proto_default_scaling_gt_pose"
# self.config['train']['dataset'] = "logs_proto_unit_scaling_gt_pose"
# self.config['train']['dataset'] = "logs_proto_default_scaling_gt_pose"
# self.config['train']['dataset'] = None
self.config['train']['dataset'] = "logs_proto_kitti_raw"
self.config['evaluate'] = {}
self.config['evaluate']['required'] = False
# self.config['evaluate']['model_lst'] = ['trained_models/new_test_caterpillar/default_scaling_gt_pose_3','trained_models/new_test_caterpillar/original_3','trained_models/new_test_caterpillar/unit_scaling_gt_pose_3']
self.config['evaluate']['model_lst'] = [
'experiments/exp_08202020-152733/trained_models/kitti_whole_3'
]
self.config['evaluate']['model_lst'].append(
'experiments/exp_08202020-155210/trained_models/kitti_left_3')
self.config['evaluate']['model_lst'].append(
'experiments/exp_08202020-161940/trained_models/kitti_right_3')
self.config['evaluate']['num_image_pairs'] = 100
self.config['evaluate']['gt_dataset_config_file'] = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'evaluation', 'kitti_whole.yaml')
# self.config['evaluate']['gt_dataset_config_file'] = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config','dense_correspondence', 'evaluation', 'gt_dataset.yaml')
self.config['experiments'] = {}
self.config['experiments']['as_experiment'] = True
# init start
self.config_path = config_path
# self.config = utils.getDictFromYamlFilename(config_path)
# configs about experiment phase
self.as_experiment = self.config['experiments']['as_experiment']
# create experiment folder if needed
if self.as_experiment:
self.current_timestamp = datetime.now().strftime("%m%d%Y-%H%M%S")
self.exp_dir = 'experiments/exp_' + self.current_timestamp
# configs about dataset generation
if not self.config['generate_dataset']['required']:
self.generate_dataset_required = False
else:
self.generate_dataset_required = True
if self.config['generate_dataset']['generate_depth']:
self.generate_depth_required = True
self.data_source = self.config['generate_dataset'][
'data_source']
self.original_data = self.config['generate_dataset'][
'original_data']
self.meta_dir = self.config['generate_dataset']['meta_dir']
self.output_target = self.config['generate_dataset'][
'output_target']
self.image_dir = self.config['generate_dataset']['image_dir']
self.output_dir = self.config['generate_dataset']['output_dir']
self.model_path = self.config['generate_dataset']['model_path']
self.scaling_method = self.config['generate_dataset'][
'scaling_method']
self.zero_masked = self.config['generate_dataset'][
'zero_masked']
self.ext = self.config['generate_dataset']['ext']
self.no_cuda = self.config['generate_dataset']['no_cuda']
else:
self.generate_depth_required = False
if self.config['generate_dataset']['replace_poses']:
self.replace_poses_required = True
self.pose_data_path = self.config['generate_dataset'][
'replace_poses']['pose_data_path']
else:
self.replace_poses_required = False
# set up datset name
self.dataset_name = 'logs_proto_'
if self.generate_depth_required:
self.dataset_name += self.scaling_method + '_'
else:
self.dataset_name += 'gt_depth_'
if self.replace_poses_required:
self.dataset_name += 'pred_pose'
else:
self.dataset_name += 'gt_pose'
if self.as_experiment:
self.dataset_name = self.exp_dir + '/' + self.dataset_name
if self.meta_dir is None:
self.meta_dir = os.path.join(self.data_source,
self.dataset_name)
# configs about training phase
if not self.config['train']['required']:
self.train_required = False
else:
self.train_required = True
self.train_dataset_config_file = self.config['train'][
'dataset_config_file']
self.train_config_file = self.config['train']['train_config_file']
self.train_logging_dir = self.config['train']['logging_dir']
self.train_num_iterations = self.config['train']['num_iterations']
self.train_dimension = self.config['train']['dimension']
self.train_dataset = ''
if self.generate_dataset_required:
self.train_dataset = self.dataset_name
if self.config['train']['dataset'] is not None:
self.train_dataset = self.config['train']['dataset']
if self.as_experiment:
self.train_logging_dir = self.exp_dir + '/trained_models'
# configs about evaluation phase
if not self.config['evaluate']['required']:
self.evaluate_required = False
else:
self.evaluate_required = True
self.eval_model_lst = self.config['evaluate']['model_lst']
self.eval_num_image_pairs = self.config['evaluate'][
'num_image_pairs'] = 100
self.eval_gt_dataset_config_file = self.config['evaluate'][
'gt_dataset_config_file']
import os
import cv2
import numpy as np
import copy
from config.params import *
import logging
os.environ['DC_SOURCE_DIR'] = DIR_PROJ
os.environ['DC_DATA_DIR'] = "{}/pdc".format(DIR_DATA)
import ipdb
st = ipdb.set_trace
# st()
import dense_correspondence_manipulation.utils.utils as utils
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
sys.path.append(dc_source_dir)
sys.path.append(os.path.join(dc_source_dir, "dense_correspondence", "correspondence_tools"))
from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset, ImageType
import dense_correspondence
from dense_correspondence.evaluation.evaluation import *
from dense_correspondence.evaluation.plotting import normalize_descriptor
from dense_correspondence.network.dense_correspondence_network import DenseCorrespondenceNetwork
import dense_correspondence_manipulation.utils.visualization as vis_utils
from dense_correspondence_manipulation.simple_pixel_correspondence_labeler.annotate_correspondences import label_colors, draw_reticle, pil_image_to_cv2, drawing_scale_config, numpy_to_cv2
def load_default_config():
default_config_file = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'stations', 'RLG_iiwa_1',
'change_detection.yaml')
config = utils.getDictFromYamlFilename(default_config_file)
return config
import dense_correspondence_manipulation.utils.utils as utils
utils.add_dense_correspondence_to_python_path()
from dense_correspondence.training.training import *
import sys
import logging
#utils.set_default_cuda_visible_devices()
# utils.set_cuda_visible_devices([0]) # use this to manually set CUDA_VISIBLE_DEVICES
from dense_correspondence.training.training import DenseCorrespondenceTraining
from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset
logging.basicConfig(level=logging.INFO)
from dense_correspondence.evaluation.evaluation import DenseCorrespondenceEvaluation
config_filename = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
'dataset', 'composite', 'toy.yaml')
config = utils.getDictFromYamlFilename(config_filename)
train_config_file = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
'training', 'toy_training.yaml')
train_config = utils.getDictFromYamlFilename(train_config_file)
dataset = SpartanDataset(config=config)
logging_dir = "/home/zhouxian/git/pytorch-dense-correspondence/pdc/trained_models/tutorials"
d = 3 # the descriptor dimension
name = "toy_hacker_%d" %(d)
train_config["training"]["logging_dir_name"] = name
train_config["training"]["logging_dir"] = logging_dir
train_config["dense_correspondence_network"]["descriptor_dimension"] = d
def load_default_config():
default_config_file = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config', 'stations',
'RLG_iiwa_1', 'change_detection.yaml')
config = utils.getDictFromYamlFilename(default_config_file)
return config
def run(data_folder, ply_binary_filename='images.ply'):
# install ply if do not already have it
os.chdir(data_folder)
vtp_filename = os.path.join(data_folder, 'images.vtp')
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
ply_to_ascii_executable = os.path.join(dc_source_dir, 'src', 'ply', 'ply2ascii')
path_to_ply = os.path.join(dc_source_dir, "src", "ply")
if not (os.path.isfile(ply_to_ascii_executable)):
os.system("cd " + path_to_ply + " && make")
correct_ply_header_file = os.path.join(dc_source_dir, 'config', 'correct_ply_header.txt')
ply_binary_full_filename = os.path.join(data_folder, ply_binary_filename)
converted_ascii_filename = os.path.join(data_folder, "converted_to_ascii.ply")
converted_ascii_modified_header_filename = os.path.join(data_folder, "converted_to_ascii_modified_header.ply")
# call ply2ascii
os.system(ply_to_ascii_executable + "<./" + ply_binary_filename + "> " + converted_ascii_filename)
# change header to be compatible with Director
# TODO: make so Director accepts other header?
line_elements_vertex = ""
with open(converted_ascii_modified_header_filename, 'w') as outfile:
with open(converted_ascii_filename) as infile:
counter = 0
for line in infile:
counter += 1
if counter == 3:
line_elements_vertex = line
break
with open(correct_ply_header_file) as infile:
counter = 0
for line in infile:
counter += 1
if counter == 4:
outfile.write(line_elements_vertex)
continue
outfile.write(line)
with open(converted_ascii_filename) as infile:
num_skip = 14
counter = 0
for line in infile:
counter += 1
if counter <= 14:
continue
outfile.write(line)
# convert to vtp
convert_ply_to_vtp_script = os.path.join(dc_source_dir, 'modules',
'dense_correspondence_manipulation', 'scripts', 'convertPlyToVtp.py')
print "converted to ascii ply format"
os.system("directorPython " + convert_ply_to_vtp_script + " " + converted_ascii_modified_header_filename)
converted_ascii_modified_header_vtp_filename = os.path.join(data_folder, "converted_to_ascii_modified_header.vtp")
print "finished convert_ply_to_vtp_script"
# clean up and rename
# os.system("rm *.ply *.freiburg")
os.rename(converted_ascii_modified_header_vtp_filename, vtp_filename)
import os
import dense_correspondence_manipulation.utils.utils as utils
CHANGE_DETECTION_CONFIG_FILE = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config', 'stations',
'RLG_iiwa_1', 'change_detection.yaml')
CHANGE_DETECTION_BACKGROUND_SUBTRACTION_CONFIG_FILE = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config', 'stations',
'RLG_iiwa_1', 'change_detection_background_subtraction.yaml')
BACKGROUND_SCENE_DATA_FOLDER = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'data_volume', 'pdc',
'logs_proto', '14_background')
DEPTH_IM_SCALE = 1000.0 # This represents that depth images are saved as uint16, where the integer value
# is depth in millimeters. So this scale just converts millimeters to meters.
DEPTH_IM_RESCALE = 4000.0 # Only for visualizaiton purposes
IMAGE_NET_MEAN = [0.485, 0.456, 0.406]
IMAGE_NET_STD_DEV = [0.229, 0.224, 0.225]
DEFAULT_IMAGE_MEAN = [
0.5573105812072754, 0.37420374155044556, 0.37020164728164673
]
DEFAULT_IMAGE_STD_DEV = [
0.24336038529872894, 0.2987397611141205, 0.31875079870224
]
DCE = DenseCorrespondenceEvaluation
for subdir in model_lst:
print("evaluate model {}".format(subdir))
start_time = time.time()
output_subdir = os.path.join(utils.get_data_dir(), output_dir, subdir.split('/')[-1])
DCE.run_evaluation_on_network(model_folder=subdir, compute_descriptor_statistics=True, cross_scene=False,
output_dir=output_subdir, num_image_pairs=num_image_pairs,dataset=gt_dataset)
end_time = time.time()
print("evaluation takes %.2f seconds" %(end_time - start_time))
if __name__ == '__main__':
args = parse_args()
model_dir = args.model_dir
output_dir = args.output_dir
print('output_dir')
print(output_dir)
num_image_pairs = args.num_image_pairs
gt_dataset_config_file = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config','dense_correspondence', 'evaluation', 'gt_dataset.yaml')
# gt_dataset_config_file = '/home/kudo/data/pdc/trained_models/new_test_caterpillar/default_scaling_gt_pose_3/dataset.yaml'
gt_dataset_config = utils.getDictFromYamlFilename(gt_dataset_config_file)
gt_dataset_config = None
abs_model_dir = utils.convert_data_relative_path_to_absolute_path(model_dir)
d = abs_model_dir
model_lst = [model_dir + '/' + o for o in os.listdir(d) if os.path.isdir(os.path.join(d,o))]
print("model list")
print(model_lst)
evaulate_model(model_lst, output_dir, num_image_pairs, gt_dataset_config)
import sys
import os
import cv2
import numpy as np
import copy
import dense_correspondence_manipulation.utils.utils as utils
import time
from PIL import Image
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
sys.path.append(dc_source_dir)
sys.path.append(
os.path.join(dc_source_dir, "dense_correspondence",
"correspondence_tools"))
from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset, ImageType
import dense_correspondence
from dense_correspondence.evaluation.evaluation import *
from dense_correspondence.evaluation.plotting import normalize_descriptor
from dense_correspondence.network.dense_correspondence_network import DenseCorrespondenceNetwork
import dense_correspondence.correspondence_tools.correspondence_finder as correspondence_finder
sys.path.append(
os.path.join(os.path.dirname(__file__),
"../simple-pixel-correspondence-labeler"))
from annotate_correspondences import label_colors, draw_reticle, pil_image_to_cv2, drawing_scale_config, numpy_to_cv2, label_colors
COLOR_RED = np.array([0, 0, 255])
COLOR_GREEN = np.array([0, 255, 0])
utils.set_default_cuda_visible_devices()
import dense_correspondence.correspondence_tools.correspondence_plotter as correspondence_plotter
from dense_correspondence.dataset.dense_correspondence_dataset_masked import ImageType
import os
import torch
import numpy as np
import argparse
#%matplotlib inline
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--data_name",
type=str,
default="caterpillar_upright.yaml")
args = parser.parse_args()
dataset_config_filename = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config',
'dense_correspondence', 'dataset', 'composite', args.data_name)
dataset_config = utils.getDictFromYamlFilename(dataset_config_filename)
dataset = SpartanDataset(debug=True, config=dataset_config)
match_type, image_a_rgb, image_b_rgb, \
matches_a, matches_b, masked_non_matches_a, \
masked_non_matches_a, non_masked_non_matches_a, \
non_masked_non_matches_b, blind_non_matches_a, \
blind_non_matches_b, metadata = dataset.get_single_object_within_scene_data()
import sys
import os
import cv2
import numpy as np
import copy
import dense_correspondence_manipulation.utils.utils as utils
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
sys.path.append(dc_source_dir)
sys.path.append(
os.path.join(dc_source_dir, "dense_correspondence",
"correspondence_tools"))
from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset, ImageType
from dense_correspondence_manipulation.simple_pixel_correspondence_labeler.annotate_correspondences import label_colors, draw_reticle, pil_image_to_cv2, drawing_scale_config
config_filename = os.path.join(
utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
'dataset', 'composite', 'caterpillar_baymax_starbot_onlymulti_front.yaml')
config = utils.getDictFromYamlFilename(config_filename)
sd = SpartanDataset(config=config)
sd.set_train_mode()
annotated_data_yaml_filename = os.path.join(os.getcwd(),
"new_annotated_pairs.yaml")
annotated_data = utils.getDictFromYamlFilename(annotated_data_yaml_filename)
index_of_pair_to_display = 0
def draw_points(img, img_points_picked):
import sys
import os
import cv2
import numpy as np
import copy
import dense_correspondence_manipulation.utils.utils as utils
dc_source_dir = utils.getDenseCorrespondenceSourceDir()
sys.path.append(dc_source_dir)
sys.path.append(os.path.join(dc_source_dir, "dense_correspondence", "correspondence_tools"))
from dense_correspondence.dataset.spartan_dataset_masked import SpartanDataset, ImageType
from dense_correspondence_manipulation.simple_pixel_correspondence_labeler.annotate_correspondences import label_colors, draw_reticle, pil_image_to_cv2, drawing_scale_config
config_filename = os.path.join(utils.getDenseCorrespondenceSourceDir(), 'config', 'dense_correspondence',
'dataset', 'composite', 'caterpillar_baymax_starbot_onlymulti_front.yaml')
config = utils.getDictFromYamlFilename(config_filename)
sd = SpartanDataset(config=config)
sd.set_train_mode()
annotated_data_yaml_filename = os.path.join(os.getcwd(), "new_annotated_pairs.yaml")
annotated_data = utils.getDictFromYamlFilename(annotated_data_yaml_filename)
index_of_pair_to_display = 0
def draw_points(img, img_points_picked):
for index, img_point in enumerate(img_points_picked):
color = label_colors[index%len(label_colors)]
draw_reticle(img, int(img_point["u"]), int(img_point["v"]), color)
def next_image_pair_from_saved():
