def show_results(img, results, dataset):
img = np.minimum(img, 1500)
img = (img - img.min()) / (img.max() - img.min())
img = np.uint8(img*255)
img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
img_show = util.draw_pose(dataset, img, results)
return img_show
def show_pose(dataset_model, dataset_image, base_dir, outputs, list_file,
save_dir, is_flip, gif):
if list_file is None:
names = util.load_names(dataset_image)
else:
with open(list_file) as f:
names = [line.strip() for line in f]
assert len(names) == outputs.shape[0]
for idx, (name, pose) in enumerate(zip(names, outputs)):
img = util.load_image(dataset_image,
os.path.join(base_dir, name),
is_flip=is_flip)
img = img.astype(np.float32)
img = (img - img.min()) / (img.max() - img.min()) * 255
img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
img = util.draw_pose(dataset_model, img, pose)
cv2.imshow('result', img / 255)
if save_dir is not None:
cv2.imwrite(os.path.join(save_dir, '{:>06d}.png'.format(idx)), img)
ch = cv2.waitKey(25)
if ch == ord('q'):
break
if gif and save_dir is not None:
os.system(
'convert -loop 0 -page +0+0 -delay 25 {0}/*.png {0}/output.gif'.
format(save_dir))
def show_results(img, results, cropped_image, dataset):
img = np.minimum(img, 1500)
img = (img - img.min()) / (img.max() - img.min())
img = np.uint8(img * 255)
# draw cropped image
img[:96, :96] = (cropped_image + 1) * 255 / 2
img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
cv2.rectangle(img, (0, 0), (96, 96), (255, 0, 0), thickness=2)
img_show = util.draw_pose(dataset, img, results)
return img_show
### this segment is only for validation
##############################################################################################
lbls = util.load_labels(dataset, phase) ### load test/train data
names = util.load_names(dataset, phase)
centers = util.load_centers(dataset, phase).astype(float)
for idx, name in enumerate(names):
if idx % 1000 == 0:
lbl = np.asarray(np.reshape(lbls[idx].split(), (21, 3)),
dtype=np.float32)
lbl, skel_camcoords = util.world2pixel(lbl, dataset)
img = util.load_image(dataset, os.path.join(root_dir, name))
img /= 1160
img *= 255
points = centers[idx]
img = util.draw_pose(dataset, img, lbl, 3, (255, 0, 0), points)
cv2.imwrite(
root_dir + 'samples/depth/' + phase + '_' + str(idx) + '.png', img)
##############################################################################################
############################# Normalize Labels #################################
### Make sure centers are calculated using get_centers.py before running this
### Also make sure you are working with original non normalized 3D joint coordinates not the normalized ones.
### After running this segment you should rename the original and give this new file the original's name so that it becomes the labels file used moving forward
################################################################################
lbls = util.load_labels(dataset, phase) # load original test/train data
centers = util.load_centers(dataset, phase).astype(float)
fx, fy, ux, uy = util.get_param(dataset)
out_file = root_dir + 'labels/fpad_' + phase + '_label_nm.txt'
lbls = np.asarray([s.split() for s in lbls], dtype=np.float32)
centers = util.load_centers(dataset,phase).astype(float)
fx, fy, ux, uy = util.get_param(dataset)
lbls = [s.split() for s in lbls]
lbls = np.reshape(np.asarray(lbls, dtype=np.float32),(-1,63))
lbls = util.transform_pose(dataset, lbls, centers, 150, fx, fy) # norm to 2D pixel
centers = np.reshape(centers,(-1,3))
for idx, name in enumerate(names):
if idx%1000 == 0:
lbl = util.pixel2world(lbls[idx], dataset) # pixel to 3D world
lbl, skel_camcoords = util.world2pixel(lbl, dataset) # back to 2d pixel from 3D world
img = util.load_image(dataset, os.path.join(root_dir, name))
points = centers[idx]
img = util.draw_pose(dataset, img, lbl, 5, (0,255,0), points)
cv2.imwrite(root_dir+'samples/rgb/from_norm/'+phase+'_'+str(idx)+'.png', img)
plt = util.plot_joints(dataset, os.path.join(root_dir, name), lbl, skel_camcoords, points)
plt.savefig(root_dir+'samples/rgb/from_norm/'+phase+'_'+str(idx)+'_plot.png')
#############################################################################################################################################
############################ Resize RGB images to match input size #################################
### Crop the RGB images using the crop size in mm and store in cropped directory
#####################################################################################################
cropped = 'cropped/'
names = util.load_names(dataset,phase)
centers = util.load_centers(dataset,phase).astype(float)
centers = np.reshape(centers, (-1,3))
cords_d = np.reshape(cords_d, (-1, 3))
cords_3d = util.pixel2world(cords_d, 'fpad')
cords_c, skel_camcoords = util.world2pixel(cords_3d, 'fpac')
cords_3d = np.reshape(cords_3d, (480, 640, -1))
img_rgbd = np.zeros((img.shape[0], img.shape[1], 4))
cords_c = np.reshape(cords_c, (480, 640, 3))
img_rgbd = image_loops.color_map(img, cimg, img_rgbd, cords_c)
img_rgbd = np.asarray(img_rgbd)
img = np.asarray(img_rgbd[:, :, :3]).copy()
pred, skel_camcoords = util.world2pixel(preds[idx], 'fpad')
label, skel_camcoords = util.world2pixel(lbls[idx], 'fpad')
points = centers[idx]
img = util.draw_pose('fpad', img, pred, 3, (0, 0, 255))
img = cv2.resize(img, (320, 240))
cv2.imwrite(root_dir + 'samples/rgbd/predictions/' + str(idx) + '.png',
img)
video_name = root_dir + 'samples/rgbd/predictions/videos/' + action + '.avi'
image_folder = root_dir + 'samples/rgbd/predictions/'
images = [img for img in os.listdir(image_folder) if img.endswith(".png")]
images = [img.replace('.png', '') for img in images]
images.sort(key=float)
images = [img + '.png' for img in images]
frame = cv2.imread(os.path.join(image_folder, images[0]))
height, width, layers = frame.shape
x1 = train_d[i, :, :, :]
if dataset in ('rgbd', 'fpac'):
x1 = np.swapaxes(x1, 2, 0)
x1 += 1
x1 *= 255
x1 /= 2
y1 = train_j[i, :]
y1 = y1.reshape(J, 3)
y1 = (y1 + 1) / 2 * 96
if dataset == 'fpad':
x1 = x1.reshape(96, 96, 1)
img = cv2.cvtColor(x1, cv2.COLOR_GRAY2RGB)
else:
img = x1[:, :, :3].copy()
if dataset == 'rgbd':
img = util.draw_pose('fpad', img, y1, 2, (0, 0, 255))
else:
img = util.draw_pose(dataset, img, y1, 2, (0, 0, 255))
img = cv2.resize(img, (200, 200))
cv2.imwrite(
root_dir + 'trial/' + dataset + '/train_' + str(i) + '.png', img)
for i in range(test_d.shape[0]):
if i % 1000 == 0:
x2 = test_d[i, :, :, :]
if dataset in ('rgbd', 'fpac'):
x2 = np.swapaxes(x2, 2, 0)
x2 += 1
x2 *= 255
x2 /= 2
y2 = test_j[i, :]