def main():
try:
# Starting OpenPose
opWrapper = op.WrapperPython()
opWrapper.configure(params)
opWrapper.start()
# Read frames on directory
imagePaths = op.get_images_on_directory(args[0].image_dir)
start = time.time()
# Process and display images
for imagePath in imagePaths:
datum = op.Datum()
imageToProcess = cv2.imread(imagePath)
datum.cvInputData = imageToProcess
opWrapper.emplaceAndPop([datum])
print("Body keypoints: \n" + str(datum.poseKeypoints))
#cv2.imshow("OpenPose 1.4.0 - Tutorial Python API", datum.cvOutputData)
cv2.imwrite("/home/lburget/Pictures/Webcam/test.jpg",
datum.cvOutputData)
end = time.time()
print("OpenPose demo successfully finished. Total time: " +
str(end - start) + " seconds")
except Exception as e:
print(e)
if key not in params: params[key] = "1"
elif "--" in curr_item and "--" not in next_item:
key = curr_item.replace('-', '')
if key not in params: params[key] = next_item
# Construct it from system arguments
# op.init_argv(args[1])
# oppython = op.OpenposePython()
# Starting OpenPose
opWrapper = op.WrapperPython()
opWrapper.configure(params)
opWrapper.start()
# Read frames on directory
imagePaths = op.get_images_on_directory(args[0].image_dir)
start = time.time()
# Process and display images
for imagePath in imagePaths:
datum = op.Datum()
imageToProcess = cv2.imread(imagePath)
datum.cvInputData = imageToProcess
opWrapper.emplaceAndPop([datum])
print("Body keypoints: \n" + str(datum.poseKeypoints))
if not args[0].no_display:
cv2.imshow("OpenPose 1.5.1 - Tutorial Python API",
datum.cvOutputData)
key = cv2.waitKey(15)
def extract_dataset_features(dataset,
isExtractRoiHogBitmap= False,
isDisplayImageWithROIs= False):
#
# extract keypoints features and HOG features from dataset
# input:
# dataset - the path of dataset for features extracting
# output:
# void - all features data are saved to disk
#
try:
# region initialization for OpenPose to run
# code from OpenPose library tutorial
# Import Openpose (Windows/Ubuntu/OSX)
dir_path = os.path.dirname(os.path.abspath(__file__))
try:
# Windows Import
if platform == "win32":
# Change these variables to point to the correct folder (Release/x64 etc.)
sys.path.append(dir_path + '/dependencies/openpose/libs');
os.environ['PATH'] = os.environ['PATH'] + ';' + dir_path + '/dependencies/openpose/dlls;'
import pyopenpose as op
else:
# Change these variables to point to the correct folder (Release/x64 etc.)
sys.path.append('../../python');
# If you run `make install` (default path is `/usr/local/python` for Ubuntu), you can also access the OpenPose/python module from there. This will install OpenPose and the python library at your desired installation path. Ensure that this is in your python path in order to use it.
# sys.path.append('/usr/local/python')
from openpose import pyopenpose as op
except ImportError as e:
print(
'Error: OpenPose library could not be found. Did you enable `BUILD_PYTHON` in CMake and have this Python script in the right folder?')
raise e
# Custom Params (refer to include/openpose/flags.hpp for more parameters)
params = dict()
params["model_folder"] = "./dependencies/openpose/models/"
params["net_resolution"] = "320x176"
# Flags
parser = argparse.ArgumentParser()
args = parser.parse_known_args()
# parse other params passed
for i in range(0, len(args[1])):
curr_item = args[1][i]
if i != len(args[1]) - 1:
next_item = args[1][i + 1]
else:
next_item = "1"
if "--" in curr_item and "--" in next_item:
key = curr_item.replace('-', '')
if key not in params: params[key] = "1"
elif "--" in curr_item and "--" not in next_item:
key = curr_item.replace('-', '')
if key not in params: params[key] = next_item
# Starting OpenPose
opWrapper = op.WrapperPython()
opWrapper.configure(params)
opWrapper.start()
# endregion
# get all training poses as sub-folders of train dataset
categories = [f.name for f in os.scandir(join(dir_path, dataset)) if f.is_dir()]
if len(categories) == 0:
parent, category = os.path.split(dataset)
dataset = parent
categories = [category]
for category in categories:
# Get key-points features and image name and size from train dataset
# declare variables hold features and labels
record_details = []
# build category path
cat_path = join(dir_path, dataset, category)
# create ROI and HOG directory
roi_and_hog_path = ""
if isExtractRoiHogBitmap:
roi_and_hog_path = join(cat_path, "RoI-HOG")
if not os.path.exists(roi_and_hog_path):
os.mkdir(roi_and_hog_path)
# get all images from the category
image_paths = op.get_images_on_directory(cat_path)
# process all images
for image_path in image_paths:
print("\nCurrent image: " + image_path + "\n")
# get image full name (with file extension)
_, image_full_name = os.path.split(image_path)
# image name without extension
image_name = image_full_name[0:image_full_name.rindex('.')]
# processed by OpenPose
datum = op.Datum()
imageToProcess = cv2.imread(image_path)
height, width, channels = imageToProcess.shape
datum.cvInputData = imageToProcess
opWrapper.emplaceAndPop([datum])
# print("Body keypoints: \n" + str(datum.poseKeypoints))
# check if exists pose key points data
if not (datum.poseKeypoints.size < 2):
# merging bounding boxes if applicable
merged_poseKeypoints = bounding_box_merging(datum.poseKeypoints,
ratio_of_intersec_thresh= 0.36,
ratio_of_distance_thresh= 2)
# assign merged_poseKeypoints to datum.poseKeypoints array
datum.poseKeypoints = merged_poseKeypoints
# an array to save mixed features (keypoints and HOG) of an image
keyPoint_HOG_feats_arrs = []
if datum.poseKeypoints.size > 1 and datum.poseKeypoints.ndim == 3:
# region extract keypoints features
keypoints_coordinates = datum.poseKeypoints[:, :, :2]
# TODO: 3. after merging with new keypoints sets, these coordinates are translated to their center and scaled by their box size -> added in keypoints features
# translate and scale keypoints by its center and box size
flatted_features = normalize_keypoint_by_its_bounding_box(keypoints_coordinates)
# image_size_in_str = str(width) + 'x' + str(height)
for row in flatted_features:
keyPoint_HOG_feats_arrs.append(row)
# endregion
# TODO: 4. RoI and HOG features are extracted -> added in HOG features
#region extract HOG features
extracted_hog_feats_arrs = \
extract_ROI_and_HOG_feature(datum,
image_name,
roi_and_hog_path,
isDisplayImageWithROIs=isDisplayImageWithROIs,
isExtractRoiHogBitmap=isExtractRoiHogBitmap)
for idx in range(0,len(extracted_hog_feats_arrs)):
keyPoint_HOG_feats_arrs[idx] = np.append(keyPoint_HOG_feats_arrs[idx],
extracted_hog_feats_arrs[idx])
# keyPoint_HOG_feats_arrs[idx] = np.append(keyPoint_HOG_feats_arrs[idx],
# [image_full_name,image_size_in_str])
#endregion
# add features data to accumulate array
for row in keyPoint_HOG_feats_arrs:
record_details.append(row)
# save merged, normalized features to disk
if len(record_details) > 0:
data_to_write = pd.DataFrame(record_details)
data_to_write.to_csv(join(cat_path, category + '.csv'))
except Exception as e:
print(e)
sys.exit(-1)
