def get_mean_and_std(dataset):
'''Compute the mean and std value of dataset.'''
param_file = os.path.join('./nets/netparams/{:s}'.format(dataset.get_name))
if not os.path.isfile(param_file):
dataloader = torch.utils.data.DataLoader(dataset,
batch_size=1,
shuffle=True,
num_workers=2)
mean = torch.zeros(3)
std = torch.zeros(3)
print('==> Computing mean and std..')
for inputs, targets in dataloader:
for i in range(3):
mean[i] += inputs[:, i, :, :].mean()
std[i] += inputs[:, i, :, :].std()
mean.div_(len(dataset))
std.div_(len(dataset))
params = {}
params['mean'] = mean
params['std'] = std
load_utils.save_json(params, param_file)
else:
params = load_utils.load_json(param_file)
mean = params['mean']
std = params['std']
return mean, std
ax.set_title(titles[idx])
rect_i = patches.Rectangle((s_bbox[0], s_bbox[1]),
s_bbox[2] - s_bbox[0],
s_bbox[3] - s_bbox[1],
linewidth=2,
edgecolor='yellow',
facecolor='none')
# Add the patch to the Axes
ax.add_patch(rect_i)
plt.show(block=False)
raw_input("Press Enter to continue...")
plt.close()
annotation_path = 'ProposalResults/ViewProposalResults-tmp.txt'
image_path_root = '/home/zwei/Dev/CVPR18_release/tmp'
image_data = load_utils.load_json(annotation_path)
image_name_list = image_data.keys()
image_name_list.sort()
for idx, image_name in enumerate(image_name_list):
if idx < 20:
continue
s_image_path = os.path.join(image_path_root, image_name)
bboxes = image_data[image_name]['bboxes']
scores = image_data[image_name]['scores']
viewBBoxes(s_image_path, bboxes, scores)
print "DEBUG"