def test_crop_tensor(self):
dataset = OTB(self.otb_dir, download=True)
padding = random.choice([None, 0, 'avg'])
out_size = random.choice([255])
print('[PyTorch-crop] padding:', padding, 'out_size:', out_size)
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = cv2.imread(img_file)
if image.ndim == 2:
image = cv2.cvtColor(image, cv2.COLOR_GRAY2RGB)
elif image.ndim == 3:
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image = torch.from_numpy(image).permute(2, 0,
1).unsqueeze(0).float()
bndbox = torch.from_numpy(anno[f, :]).float()
center = bndbox[:2] + bndbox[2:] / 2
patch = crop_tensor(image,
center,
bndbox[2:],
padding=padding,
out_size=out_size)
patch = patch.squeeze().permute(1, 2,
0).cpu().numpy().astype(np.uint8)
show_frame(patch, fig_n=1, pause=0.1)
def test_pairwise(self):
base_dataset = OTB(self.otb_dir, download=True)
frame_range = random.choice([0, 1, 100])
causal = random.choice([True, False])
subset = random.choice(['train', 'val'])
return_index = random.choice([True, False])
rand_choice = random.choice([True, False])
dataset = Pairwise(
base_dataset, pairs_per_video=1, frame_range=frame_range,
causal=causal, subset=subset, return_index=return_index,
rand_choice=rand_choice)
self.assertGreater(len(dataset), 0)
for i, item in enumerate(dataset):
img_z, img_x, bndbox_z, bndbox_x = \
item[0], item[1], item[2], item[3]
if return_index:
print('rand_z:', item[4], '\trand_x:', item[5])
self.assertEqual(img_z.mode, 'RGB')
self.assertEqual(img_x.mode, 'RGB')
self.assertEqual(bndbox_z.shape, (4,))
self.assertEqual(bndbox_x.shape, (4,))
if self.visualize:
item = random.choice(dataset)
img_z, img_x, bndbox_z, bndbox_x = \
item[0], item[1], item[2], item[3]
if return_index:
print('rand_z:', item[4], '\trand_x:', item[5])
show_frame(img_z, bndbox_z, fig_n=1, pause=1)
show_frame(img_x, bndbox_x, fig_n=2, pause=1)
def test_transform_goturn(self):
base_dataset = VOT(self.vot_dir, return_rect=True, download=True)
transform = TransformGOTURN()
dataset = Pairwise(base_dataset,
transform,
pairs_per_video=1,
frame_range=1,
causal=True)
self.assertGreater(len(dataset), 0)
for crop_z, crop_x, labels in dataset:
self.assertEqual(crop_z.size(), crop_x.size())
if self.visualize:
for t in range(10):
crop_z, crop_x, labels = random.choice(dataset)
mean_color = torch.tensor(transform.mean_color).float().view(
3, 1, 1)
crop_z = F.to_pil_image((crop_z + mean_color) / 255.0)
crop_x = F.to_pil_image((crop_x + mean_color) / 255.0)
labels = labels.cpu().numpy()
labels *= transform.out_size / transform.label_scale_factor
bndbox = np.concatenate([labels[:2], labels[2:] - labels[:2]])
show_frame(crop_x, bndbox, fig_n=1, pause=1)
def test_pad_pil(self):
dataset = OTB(self.otb_dir, download=True)
npad = random.choice([0, 10, 50])
padding = random.choice([None, 0, 'avg'])
print('[PIL-pad] padding:', padding, 'npad:', npad)
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = Image.open(img_file)
image = pad_pil(image, npad, padding=padding)
show_frame(image, fig_n=1)
def test_load(self):
dataset = VOT(self.vot_dir, return_rect=True)
self.assertGreater(len(dataset), 0)
for img_files, anno in dataset:
self.assertGreater(len(img_files), 0)
self.assertEqual(len(img_files), len(anno))
if self.visualize:
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = Image.open(img_file)
show_frame(image, anno[f, :])
def test_transform_siamfc(self):
base_dataset = VOT(self.vot_dir, return_rect=True, download=True)
transform = TransformSiamFC(stats_path=self.stats_path)
dataset = Pairwise(base_dataset, transform=transform, subset='train')
self.assertGreater(len(dataset), 0)
for crop_z, crop_x, labels, weights in dataset:
self.assertAlmostEqual(weights[labels == 1].sum().item(),
weights[labels == 0].sum().item())
self.assertAlmostEqual(weights.sum().item(),
labels[labels >= 0].numel())
self.assertEqual(
weights[labels == transform.ignore_label].sum().item(), 0)
if self.visualize:
crop_z, crop_x, labels, weights = random.choice(dataset)
crop_z = F.to_pil_image(crop_z / 255.0)
crop_x = F.to_pil_image(crop_x / 255.0)
labels = self._rescale(labels.cpu().squeeze().numpy())
weights = self._rescale(weights.cpu().squeeze().numpy())
bndbox_z = np.array([31, 31, 64, 64])
bndbox_x = np.array([95, 95, 64, 64])
show_frame(crop_z, bndbox_z, fig_n=1, pause=1)
show_frame(crop_x, bndbox_x, fig_n=2, pause=1)
show_frame(labels, fig_n=3, pause=1, cmap='hot')
show_frame(weights, fig_n=4, pause=5, cmap='hot')
def test_imagenet_obj(self):
subset = random.choice(['train', 'val'])
dataset = ImageNetObject(self.obj_dir, subset=subset, return_rect=True)
self.assertGreater(len(dataset), 0)
for i in range(10):
img_file, bndbox = dataset[i]
self.assertTrue(os.path.isfile(img_file))
self.assertTrue(len(bndbox) == 4)
if self.visualize:
img_file, bndbox = random.choice(dataset)
image = Image.open(img_file)
show_frame(image, bndbox, fig_n=1, pause=0.1)
def test_imagenet_vid(self):
dataset = ImageNetVID(self.vid_dir, return_rect=True)
self.assertGreater(len(dataset), 0)
for i in range(10):
img_files, anno = random.choice(dataset)
self.assertGreater(len(img_files), 0)
self.assertEqual(len(img_files), len(anno))
if self.visualize:
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = Image.open(img_file)
show_frame(image, anno[f, :])
def test_pad_array(self):
dataset = OTB(self.otb_dir, download=True)
npad = random.choice([0, 10, 50])
padding = random.choice([None, 0, 'avg'])
print('[cv2-pad] padding:', padding, 'npad:', npad)
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = cv2.imread(img_file)
if image.ndim == 2:
image = cv2.cvtColor(image, cv2.COLOR_GRAY2RGB)
elif image.ndim == 3:
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image = pad_array(image, npad, padding=padding)
show_frame(image[:, :, ::-1], fig_n=1)
def test_crop_pil(self):
dataset = OTB(self.otb_dir, download=True)
padding = random.choice([None, 0, 'avg'])
out_size = random.choice([None, 255])
print('[PIL-crop] padding:', padding, 'out_size:', out_size)
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = Image.open(img_file)
bndbox = anno[f, :]
center = bndbox[:2] + bndbox[2:] / 2
patch = crop_pil(image,
center,
bndbox[2:],
padding=padding,
out_size=out_size)
show_frame(patch, fig_n=2, pause=0.1)
def test_resize_tensor(self):
dataset = OTB(self.otb_dir, download=True)
out_size = random.choice([30, 100, 255])
print('[PyTorch-resize]:', out_size)
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = cv2.imread(img_file)
if image.ndim == 2:
image = cv2.cvtColor(image, cv2.COLOR_GRAY2RGB)
elif image.ndim == 3:
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image = torch.from_numpy(image).permute(2, 0,
1).unsqueeze(0).float()
image = resize_tensor(image, out_size)
image = image.squeeze().permute(1, 2, 0).numpy().astype(np.uint8)
show_frame(image, fig_n=2, pause=0.1)
def test_crop_array(self):
dataset = OTB(self.otb_dir, download=True)
padding = random.choice([None, 0, 'avg'])
out_size = random.choice([None, 255])
print('[cv2-crop] padding:', padding, 'out_size:', out_size)
img_files, anno = random.choice(dataset)
for f, img_file in enumerate(img_files):
image = cv2.imread(img_file)
if image.ndim == 2:
image = cv2.cvtColor(image, cv2.COLOR_GRAY2RGB)
elif image.ndim == 3:
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
bndbox = anno[f, :]
center = bndbox[:2] + bndbox[2:] / 2
patch = crop_array(image,
center,
bndbox[2:],
padding=padding,
out_size=out_size)
show_frame(patch, fig_n=2, pause=0.1)