def colors_in_image(uri) -> Set:
"""Returns set of colors that appears in the image.
Args:
uri: path to the image
Returns:
Set: colors that appears in the image
"""
image = imread(uri, rootpath=args.in_dir)
colors = np.unique(image.reshape(-1, image.shape[-1]), axis=0)
result = {tuple(row) for row in colors.tolist()} # np.array to hashable
return result
def preprocess(self, image_path: Path):
image = imread(image_path, rootpath=str(self.in_dir))
heigth, width = image.shape[:2]
mask = np.zeros((heigth, width, len(self.index2color)), dtype=np.uint8)
for index, color in self.index2color.items():
mask[np.all((image == color), axis=-1), index] = 255
target_path = self.out_dir / f"{image_path.stem}.tiff"
target_path.parent.mkdir(parents=True, exist_ok=True)
mimwrite_with_meta(
target_path, np.dsplit(mask, mask.shape[2]), {"compress": 9}
)
def _load_data(data_dir: Path) -> Tuple[torch.Tensor, torch.Tensor]:
"""
Load data from train directory of the dataset.
Parse names of images to get person id as labels.
Args:
data_dir: path to directory that contains training data
Returns:
images for training and their labels
"""
filenames = list(data_dir.glob("*.jpg"))
data = torch.from_numpy(np.array([imread(filename) for filename in filenames])).float()
targets = torch.from_numpy(
np.array([int(filename.name.split("_")[0]) for filename in filenames])
)
return data, targets
def _load_data(data_dir: Path) -> Tuple[torch.Tensor, Iterable, Iterable]:
"""
Load data from directory.
Parse names of images to get person ids as labels and camera ids.
Args:
data_dir: path to directory that contains data
Returns:
images, their labels and ids of the cameras that made the photos
"""
# Gallery dataset contains good, junk and distractor images;
# junk ones (marked as -1) should be neglected during evaluation.
filenames = list(data_dir.glob("[!-]*.jpg"))
data = torch.from_numpy(np.array([imread(filename) for filename in filenames])).float()
pids = np.array([int(filename.name.split("_")[0]) for filename in filenames])
cids = np.array([int(filename.name.split("_")[1][1:2]) for filename in filenames])
return data, pids, cids
def colors_in_image(uri) -> Set:
image = imread(uri, rootpath=args.in_dir)
colors = np.unique(image.reshape(-1, image.shape[-1]), axis=0)
result = {tuple(row) for row in colors.tolist()} # np.array to hashable
return result