charles/monoloco
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Small fixes

Browse Source
This commit is contained in:
charlesbvll 2021-03-28 17:00:53 +02:00
parent 549026513a
commit 6ca23a8f9c
2 changed files with 41 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
monoloco/activity.py
View File

@ -22,9 +22,11 @@ def social_interactions(idx, centers, angles, dds, stds=None, social_distance=Fa
# A) Check whether people are close together # A) Check whether people are close together
xx = centers[idx][0] xx = centers[idx][0]
zz = centers[idx][1] zz = centers[idx][1]
distances = [math.sqrt((xx - centers[i][0]) ** 2 + (zz - centers[i][1]) ** 2) for i, _ in enumerate(centers)] distances = [math.sqrt((xx - centers[i][0]) ** 2 + (zz - centers[i][1]) ** 2)
for i, _ in enumerate(centers)]
sorted_idxs = np.argsort(distances) sorted_idxs = np.argsort(distances)
indices = [idx_t for idx_t in sorted_idxs[1:] if distances[idx_t] <= threshold_dist] indices = [idx_t for idx_t in sorted_idxs[1:]
if distances[idx_t] <= threshold_dist]
# B) Check whether people are looking inwards and whether there are no intrusions # B) Check whether people are looking inwards and whether there are no intrusions
# Deterministic # Deterministic
@ -72,17 +74,14 @@ def is_raising_hand(keypoint):
l_hand = 9 l_hand = 9
r_shoulder = 6 r_shoulder = 6
r_hand = 10 r_hand = 10
l_ear = 3
r_ear = 4
h_offset = 20
if keypoint[1][l_hand] < keypoint[1][l_shoulder] and keypoint[1][r_hand] < keypoint[1][r_shoulder]: if keypoint[1][l_hand] < keypoint[1][l_shoulder] and keypoint[1][r_hand] < keypoint[1][r_shoulder]:
return 'both' return 'both'
if keypoint[1][l_hand] < keypoint[1][l_shoulder]: if keypoint[1][l_hand] < keypoint[1][l_shoulder]:
return 'left' return 'left'
if keypoint[1][r_hand] < keypoint[1][r_shoulder]: if keypoint[1][r_hand] < keypoint[1][r_shoulder]:
return 'right' return 'right'
return 'none' return 'none'
@ -96,7 +95,8 @@ def check_f_formations(idx, idx_t, centers, angles, radii, social_distance=False
""" """
# Extract centers and angles # Extract centers and angles
other_centers = np.array([cent for l, cent in enumerate(centers) if l not in (idx, idx_t)]) other_centers = np.array(
[cent for l, cent in enumerate(centers) if l not in (idx, idx_t)])
theta0 = angles[idx] theta0 = angles[idx]
theta1 = angles[idx_t] theta1 = angles[idx_t]
@ -115,15 +115,18 @@ def check_f_formations(idx, idx_t, centers, angles, radii, social_distance=False
# 1) Verify they are looking inwards. # 1) Verify they are looking inwards.
# The distance between mus and the center should be less wrt the original position and the center # The distance between mus and the center should be less wrt the original position and the center
d_new = np.linalg.norm(mu_0 - mu_1) / 2 if social_distance else np.linalg.norm(mu_0 - mu_1) d_new = np.linalg.norm(
mu_0 - mu_1) / 2 if social_distance else np.linalg.norm(mu_0 - mu_1)
d_0 = np.linalg.norm(x_0 - o_c) d_0 = np.linalg.norm(x_0 - o_c)
d_1 = np.linalg.norm(x_1 - o_c) d_1 = np.linalg.norm(x_1 - o_c)
# 2) Verify no intrusion for third parties # 2) Verify no intrusion for third parties
if other_centers.size: if other_centers.size:
other_distances = np.linalg.norm(other_centers - o_c.reshape(1, -1), axis=1) other_distances = np.linalg.norm(
other_centers - o_c.reshape(1, -1), axis=1)
else: else:
other_distances = 100 * np.ones((1, 1)) # Condition verified if no other people # Condition verified if no other people
other_distances = 100 * np.ones((1, 1))
# Binary Classification # Binary Classification
# if np.min(other_distances) > radius: # Ablation without orientation # if np.min(other_distances) > radius: # Ablation without orientation
@ -163,8 +166,10 @@ def show_activities(args, image_t, output_path, annotations, dic_out):
show=args.show, show=args.show,
fig_width=10, fig_width=10,
dpi_factor=1.0) as ax: dpi_factor=1.0) as ax:
keypoint_painter.keypoints(ax, keypoint_sets, colors=colors, raise_hand=r_h) keypoint_painter.keypoints(
draw_orientation(ax, uv_centers, sizes, angles, colors, mode='front') ax, keypoint_sets, colors=colors, raise_hand=r_h)
draw_orientation(ax, uv_centers, sizes,
angles, colors, mode='front')
if 'bird' in args.output_types: if 'bird' in args.output_types:
z_max = min(args.z_max, 4 + max([el[1] for el in xz_centers])) z_max = min(args.z_max, 4 + max([el[1] for el in xz_centers]))

33
monoloco/predict.py
View File

@ -46,7 +46,8 @@ def get_torch_checkpoints_dir():
def download_checkpoints(args): def download_checkpoints(args):
torch_dir = get_torch_checkpoints_dir() torch_dir = get_torch_checkpoints_dir()
pifpaf_model = os.path.join(torch_dir, 'shufflenetv2k30-201104-224654-cocokp-d75ed641.pkl') pifpaf_model = os.path.join(
torch_dir, 'shufflenetv2k30-201104-224654-cocokp-d75ed641.pkl')
dic_models = {'keypoints': pifpaf_model} dic_models = {'keypoints': pifpaf_model}
if not os.path.exists(pifpaf_model): if not os.path.exists(pifpaf_model):
import gdown import gdown
@ -154,14 +155,16 @@ def predict(args):
# data # data
data = datasets.ImageList(args.images, preprocess=preprocess) data = datasets.ImageList(args.images, preprocess=preprocess)
if args.mode == 'stereo': if args.mode == 'stereo':
assert len(data.image_paths) % 2 == 0, "Odd number of images in a stereo setting" assert len(
data.image_paths) % 2 == 0, "Odd number of images in a stereo setting"
data_loader = torch.utils.data.DataLoader( data_loader = torch.utils.data.DataLoader(
data, batch_size=args.batch_size, shuffle=False, data, batch_size=args.batch_size, shuffle=False,
pin_memory=False, collate_fn=datasets.collate_images_anns_meta) pin_memory=False, collate_fn=datasets.collate_images_anns_meta)
for batch_i, (image_tensors_batch, _, meta_batch) in enumerate(data_loader): for batch_i, (image_tensors_batch, _, meta_batch) in enumerate(data_loader):
pred_batch = processor.batch(pifpaf_model, image_tensors_batch, device=args.device) pred_batch = processor.batch(
pifpaf_model, image_tensors_batch, device=args.device)
# unbatch (only for MonStereo) # unbatch (only for MonStereo)
for idx, (pred, meta) in enumerate(zip(pred_batch, meta_batch)): for idx, (pred, meta) in enumerate(zip(pred_batch, meta_batch)):
@ -183,7 +186,8 @@ def predict(args):
output_path = os.path.join(splits[0], 'out_' + splits[1]) output_path = os.path.join(splits[0], 'out_' + splits[1])
else: else:
file_name = os.path.basename(meta['file_name']) file_name = os.path.basename(meta['file_name'])
output_path = os.path.join(args.output_directory, 'out_' + file_name) output_path = os.path.join(
args.output_directory, 'out_' + file_name)
im_name = os.path.basename(meta['file_name']) im_name = os.path.basename(meta['file_name'])
print(f'{batch_i} image {im_name} saved as {output_path}') print(f'{batch_i} image {im_name} saved as {output_path}')
@ -195,15 +199,18 @@ def predict(args):
# 3D Predictions # 3D Predictions
if args.mode != 'keypoints': if args.mode != 'keypoints':
im_size = (cpu_image.size[0], cpu_image.size[1]) # Original im_size = (cpu_image.size[0], cpu_image.size[1]) # Original
kk, dic_gt = factory_for_gt(im_size, focal_length=args.focal, name=im_name, path_gt=args.path_gt) kk, dic_gt = factory_for_gt(
im_size, focal_length=args.focal, name=im_name, path_gt=args.path_gt)
# Preprocess pifpaf outputs and run monoloco # Preprocess pifpaf outputs and run monoloco
boxes, keypoints = preprocess_pifpaf(pifpaf_outs['left'], im_size, enlarge_boxes=False) boxes, keypoints = preprocess_pifpaf(
pifpaf_outs['left'], im_size, enlarge_boxes=False)
if args.mode == 'mono': if args.mode == 'mono':
LOG.info("Prediction with MonoLoco++") LOG.info("Prediction with MonoLoco++")
dic_out = net.forward(keypoints, kk) dic_out = net.forward(keypoints, kk)
dic_out = net.post_process(dic_out, boxes, keypoints, kk, dic_gt) dic_out = net.post_process(
dic_out, boxes, keypoints, kk, dic_gt)
if args.social_distance or (args.activities and 'social_distance' in args.activities): if args.social_distance or (args.activities and 'social_distance' in args.activities):
dic_out = net.social_distance(dic_out, args) dic_out = net.social_distance(dic_out, args)
if args.activities and 'raise_hand' in args.activities: if args.activities and 'raise_hand' in args.activities:
@ -211,9 +218,11 @@ def predict(args):
else: else:
LOG.info("Prediction with MonStereo") LOG.info("Prediction with MonStereo")
boxes_r, keypoints_r = preprocess_pifpaf(pifpaf_outs['right'], im_size) boxes_r, keypoints_r = preprocess_pifpaf(
pifpaf_outs['right'], im_size)
dic_out = net.forward(keypoints, kk, keypoints_r=keypoints_r) dic_out = net.forward(keypoints, kk, keypoints_r=keypoints_r)
dic_out = net.post_process(dic_out, boxes, keypoints, kk, dic_gt) dic_out = net.post_process(
dic_out, boxes, keypoints, kk, dic_gt)
else: else:
dic_out = defaultdict(list) dic_out = defaultdict(list)
@ -242,7 +251,8 @@ def factory_outputs(args, pifpaf_outs, dic_out, output_path, kk=None):
elif any((xx in args.output_types for xx in ['front', 'bird', 'multi'])): elif any((xx in args.output_types for xx in ['front', 'bird', 'multi'])):
LOG.info(output_path) LOG.info(output_path)
if args.activities: if args.activities:
show_activities(args, pifpaf_outs['image'], output_path, pifpaf_outs['left'], dic_out) show_activities(
args, pifpaf_outs['image'], output_path, pifpaf_outs['left'], dic_out)
else: else:
printer = Printer(pifpaf_outs['image'], output_path, kk, args) printer = Printer(pifpaf_outs['image'], output_path, kk, args)
figures, axes = printer.factory_axes(dic_out) figures, axes = printer.factory_axes(dic_out)
@ -253,4 +263,5 @@ def factory_outputs(args, pifpaf_outs, dic_out, output_path, kk=None):
json.dump(dic_out, ff) json.dump(dic_out, ff)
else: else:
LOG.info("No output saved, please select one among front, bird, multi, or pifpaf options") LOG.info(
"No output saved, please select one among front, bird, multi, or pifpaf options")