if np.isfinite(chunk).all():
output[subject][name] = np.array(
chunk, dtype='float32') / 1000
output_2d[subject][name] = list(
np.array(all_chunks_2d[i],
dtype='float32').transpose(1, 0, 2, 3))
chunk_indices.append(
(current_index, next_index, np.isfinite(chunk).all(),
split, name))
current_index = next_index
assert current_index == index[subject][canonical_name][1]
if canonical_name not in frame_mapping[subject_name]:
frame_mapping[subject_name][canonical_name] = []
frame_mapping[subject_name][canonical_name] += chunk_indices
metadata = suggest_metadata('humaneva' + str(num_joints))
output_filename = 'data_3d_' + metadata['layout_name']
output_prefix_2d = 'data_2d_' + metadata['layout_name'] + '_'
if args.convert_3d:
print('Saving...')
np.savez_compressed(output_filename, positions_3d=output)
np.savez_compressed(output_prefix_2d + 'gt',
positions_2d=output_2d,
metadata=metadata)
print('Done.')
else:
print('Please specify the dataset source')
exit(0)
type=str,
metavar='PATH',
help='input path to 2D detections')
parser.add_argument(
'-o',
'--output',
default='detectron_pt_coco',
type=str,
metavar='PATH',
help='output suffix for 2D detections (e.g. detectron_pt_coco)')
args = parser.parse_args()
if not args.input:
print('Please specify the input directory')
exit(0)
# according to output name,generate some format. we use detectron
import_func = suggest_pose_importer('detectron_pt_coco')
metadata = suggest_metadata('detectron_pt_coco')
print('Parsing 2D detections from', args.input)
keypoints = import_func(args.input)
output = keypoints.astype(np.float32)
# 生成的数据用于后面的3D检测
np.savez_compressed(output_prefix_2d + 'test' + args.output,
positions_2d=output,
metadata=metadata)
print('npz name is ', output_prefix_2d + 'test' + args.output)
current_index = index[subject][canonical_name][0]
chunk_indices = []
for i, chunk in enumerate(all_chunks):
next_index = current_index + len(chunk)
name = canonical_name + ' chunk' + str(i)
if np.isfinite(chunk).all():
output[subject][name] = np.array(chunk, dtype='float32') / 1000
output_2d[subject][name] = list(np.array(all_chunks_2d[i], dtype='float32').transpose(1, 0, 2, 3))
chunk_indices.append((current_index, next_index, np.isfinite(chunk).all(), split, name))
current_index = next_index
assert current_index == index[subject][canonical_name][1]
if canonical_name not in frame_mapping[subject_name]:
frame_mapping[subject_name][canonical_name] = []
frame_mapping[subject_name][canonical_name] += chunk_indices
metadata = suggest_metadata('humaneva' + str(num_joints))
output_filename = 'data_3d_' + metadata['layout_name']
output_prefix_2d = 'data_2d_' + metadata['layout_name'] + '_'
if args.convert_3d:
print('Saving...')
np.savez_compressed(output_filename, positions_3d=output)
np.savez_compressed(output_prefix_2d + 'gt', positions_2d=output_2d, metadata=metadata)
print('Done.')
else:
print('Please specify the dataset source')
exit(0)
if args.convert_2d:
if not args.output:
parser = argparse.ArgumentParser(description='Custom dataset creator')
parser.add_argument('-i', '--input', type=str, default='', metavar='PATH', help='detections directory')
parser.add_argument('-o', '--output', type=str, default='', metavar='PATH', help='output suffix for 2D detections')
args = parser.parse_args()
if not args.input:
print('Please specify the input directory')
exit(0)
if not args.output:
print('Please specify an output suffix (e.g. detectron_pt_coco)')
exit(0)
print('Parsing 2D detections from', args.input)
metadata = suggest_metadata('coco')
metadata['video_metadata'] = {}
output = {}
file_list = glob(args.input + '/*.npz')
for f in file_list:
canonical_name = os.path.splitext(os.path.basename(f))[0]
data, video_metadata = decode(f)
output[canonical_name] = {}
output[canonical_name]['custom'] = [data[0]['keypoints'].astype('float32')]
metadata['video_metadata'][canonical_name] = video_metadata
print('Saving...')
np.savez_compressed(output_prefix_2d + args.output, positions_2d=output, metadata=metadata)
print('Done.')
parser.add_argument('-i', '--input', default='', type=str, metavar='PATH', help='input path to 2D detections')
parser.add_argument('-o', '--output', default='', type=str, metavar='PATH', help='output suffix for 2D detections (e.g. detectron_pt_coco)')
args = parser.parse_args()
if not args.input:
print('Please specify the input directory')
exit(0)
if not args.output:
print('Please specify an output suffix (e.g. detectron_pt_coco)')
exit(0)
import_func = suggest_pose_importer(args.output)
metadata = suggest_metadata(args.output)
print('Parsing 2D detections from', args.input)
output = {}
file_list = glob(args.input + '/S*/*.mp4.npz')
for f in file_list:
path, fname = os.path.split(f)
subject = os.path.basename(path)
assert subject.startswith('S'), subject + ' does not look like a subject directory'
if '_ALL' in fname:
continue
m = re.search('(.*)\\.([0-9]+)\\.mp4\\.npz', fname)
action = m.group(1)
parser.add_argument('-i', '--input', default='', type=str, metavar='PATH', help='input path to 2D detections')
parser.add_argument('-o', '--output', default='', type=str, metavar='PATH', help='output suffix for 2D detections (e.g. detectron_pt_coco)')
args = parser.parse_args()
if not args.input:
print('Please specify the input directory')
exit(0)
if not args.output:
print('Please specify an output suffix (e.g. detectron_pt_coco)')
exit(0)
import_func = suggest_pose_importer(args.output)
metadata = suggest_metadata(args.output)
print('Parsing 2D detections from', args.input)
output = {}
file_list = glob(args.input + '/S*/*.mp4.npz')
for f in file_list:
path, fname = os.path.split(f)
subject = os.path.basename(path)
assert subject.startswith('S'), subject + ' does not look like a subject directory'
if '_ALL' in fname:
continue
m = re.search('(.*)\\.([0-9]+)\\.mp4\\.npz', fname)
action = m.group(1)
def main(args):
#if they passed a directory, get all images
if os.path.isdir(args.im_or_folder):
im_list = glob.iglob(args.im_or_folder + '/*.' + args.image_ext)
#otherwise just used the one passed image
else:
im_list = [args.im_or_folder]
#gets some suggested COCO metadata like keypoint symmetries
metadata = suggest_metadata('coco')
'''NOTES
{} defines an empty dict
[] defines an empty list
These are fundamentally different types. A dict is an associative array, a list is a standard array with integral indices.'''
#set video metadata to empty dict
metadata['video_metadata'] = {}
#intialize final output dict to empty dict
output = {}
#for each video passed
for video_name in im_list:
#get absolute path of output directory
#out_name = os.path.join(args.output_dir, os.path.basename(video_name))
out_name = os.path.join(args.output_dir, "data_2d_custom_myvideos")
#print name of video that we're processing
print('Processing {}'.format(video_name))
#intialize arrays to empty
boxes = []
segments = []
keypoints = []
#iterate over each frame in the vid
for frame_i, im in enumerate(
read_video(video_name)
): #enumerate creates array of pairs--index paired with the video name
#get the time
t = time.time()
#run the inference
kps = posenet.estimate_pose(im)
#print how long it took to process frame
print('Frame {} processed in {:.3f}s'.format(
frame_i,
time.time() - t))
has_bbox = False
'''
if kps.has('pred_boxes'):
bbox_tensor = outputs.pred_boxes.tensor.numpy()
if len(bbox_tensor) > 0:
has_bbox = True
scores = outputs.scores.numpy()[:, None]
bbox_tensor = np.concatenate((bbox_tensor, scores), axis=1)'''
'''
else:
print("No keypts or boxes found in frame, setting keypts and boxes to empty array")
kps = []
bbox_tensor = []'''
#add keypoints (excluding confidence checks) for this frame to array for whole video
keypoints.append(kps[:, :2])
#store video resolution in metadata
this_vid_metadata = {
'w': im.shape[1],
'h': im.shape[0],
}
#get name of video we're processing
canonical_name = video_name
#intialize value keyed at [canonical_name] to empty dict
output[canonical_name] = {}
#add keypoint locations to output array
output[canonical_name]['custom'] = [np.array(keypoints)]
#add video resolution metadata to output array
metadata['video_metadata'][canonical_name] = this_vid_metadata
#print(keypoints)
#export npz file containing all data retrieved by inference
np.savez_compressed(out_name, positions_2d=output, metadata=metadata)