from functools import cached_property
from pathlib import Path
from typing import Dict, List, Optional, Sequence, Union, overload
import albumentations
import cv2
import numpy as np
import tifffile
from skimage.util import img_as_float32
from ..base import MaskDataset
from ..types import BundledPath
from ..utils import bundle_list, stack_channels, stack_channels_to_rgb
[docs]class BBBC007(MaskDataset):
"""Drosophila Kc167 cells
Outline annotation
Images were acquired using a motorized Zeiss Axioplan 2 and a Axiocam MRm
camera, and are provided courtesy of the laboratory of David Sabatini at the
Whitehead Institute for Biomedical Research. Each image is roughly 512 x 512
pixels, with cells roughly 25 pixels in dimeter, and 80 cells per image on
average. The two channels (DNA and actin) of each image are stored in
separate gray-scale 8-bit TIFF files.
Parameters
----------
root_dir : str
Path to root directory
output : {'both', 'image', 'mask'}, default: 'both'
Change outputs. 'both' returns {'image': image, 'mask': mask}.
transforms : albumentations.Compose, optional
An instance of Compose (albumentations pkg) that defines augmentation in
sequence.
num_samples : int, optional
Useful when ``transforms`` is set. Define the total length of the
dataset. If it is set, it overwrites ``__len__``.
grayscale : bool, default: False
Convert images to grayscale
grayscale_mode : {'equal', 'cv2', Sequence[float]}, default: 'equal'
How to convert to grayscale. If set to 'cv2', it follows opencv
implementation. Else if set to 'equal', it sums up values along channel
axis, then divides it by the number of expected channels.
image_ch : {'DNA', 'actin'}, default: ('DNA', 'actin')
Which channel(s) to load as image. Make sure to give it as a Sequence
when choose a single channel. Name matches to `anno_ch`.
anno_ch : {'DNA', 'actin'}, default: ('DNA',)
Which channel(s) to load as annotation. Make sure to give it as a
Sequence when choose a single channel.
Notes
-----
- [4, 5, 11, 14, 15] have 3 channels but they are just all gray scale
images. Extra work is required in get_image().
References
----------
.. [1] Jones et al., in the Proceedings of the ICCV Workshop on Computer
Vision for Biomedical Image Applications (CVBIA), 2005.
.. [2] https://bbbc.broadinstitute.org/BBBC007
See Also
--------
MaskDataset : Super class
Dataset : Base class
DatasetInterface : Interface
"""
# Dataset's acronym
acronym = 'BBBC007'
def __init__(
self,
root_dir: str,
*,
output: str = 'both',
transforms: Optional[albumentations.Compose] = None,
num_samples: Optional[int] = None,
grayscale: bool = False,
grayscale_mode: Union[str, Sequence[float]] = 'equal',
# specific to this dataset
image_ch: Sequence[str] = ('DNA', 'actin',),
anno_ch: Sequence[str] = ('DNA',),
**kwargs
):
self._root_dir = root_dir
self._output = output
self._transforms = transforms
self._num_samples = num_samples
self._grayscale = grayscale
self._grayscale_mode = grayscale_mode
self.image_ch = image_ch
self.anno_ch = anno_ch
if not any([ch in ('DNA', 'actin') for ch in image_ch]):
raise ValueError("Set `anno_ch` in ('nuclei', 'cells') in sequence")
if not any([ch in ('DNA', 'actin') for ch in anno_ch]):
raise ValueError("Set `anno_ch` in ('nuclei', 'cells') in sequence")
@staticmethod
def _imread_handler(p: Path) -> np.ndarray:
"""Handle irregular images by wrapping tifffile.imread
Normally two images in a pair have gray scale and only have one channel.
This means that each image array has shape of (height, width). But there
are some outliers.
For example a pair of images below has 3 channels with all having the
same value (height, width, 3).
['BBBC007_v1_images/f113/AS_09125_040701150004_A02f00d0.tif',
'BBBC007_v1_images/f113/AS_09125_040701150004_A02f00d1.tif']
6 pairs out of 16 have this issue and this wrapper resolves it.
"""
img = tifffile.imread(p)
if img.shape[-1] == 3:
return img[..., 0]
return img
[docs] def get_image(self, p: Union[Path, BundledPath]) -> np.ndarray:
if isinstance(p, Path):
img = self._imread_handler(p)
img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)
else:
img = stack_channels_to_rgb(self._imread_handler, p)
return img_as_float32(img)
[docs] def get_mask(self, p: Union[Path, BundledPath]) -> np.ndarray:
if isinstance(p, Path):
mask = tifffile.imread(p)
else:
mask = stack_channels(tifffile.imread, p)
# dtype=bool originally and bool is not well handled by albumentations
return 255 * mask.astype(np.uint8)
@cached_property
def file_list(self) -> Union[List[Path], List[BundledPath]]:
file_list: Union[List[Path], List[List[Path]]]
root_dir = self.root_dir
parent = 'BBBC007_v1_images'
_file_list = sorted(root_dir.glob(f'{parent}/*/*.tif'))
if len(ch := self.image_ch) == 1:
if ch[0] == 'DNA':
file_list = _file_list[::2]
elif ch[0] == 'actin':
file_list = _file_list[1::2]
else:
raise ValueError("Set `image_ch` in ('DNA', 'actin')")
elif len(ch) == 2:
file_list = bundle_list(_file_list, 2)
else:
raise ValueError("Set `image_ch` in ('DNA', 'actin') or all")
return file_list
@cached_property
def anno_dict(self) -> Union[Dict[int, Path], Dict[int, BundledPath]]:
root_dir = self.root_dir
parent = 'BBBC007_v1_outlines'
_anno_list = sorted(root_dir.glob(f'{parent}/*/*.tif'))
if len(ch := self.anno_ch) == 1:
if ch[0] == 'DNA':
anno_list = _anno_list[::2]
elif ch[0] == 'actin':
anno_list = _anno_list[1::2]
return dict((k, v) for k, v in enumerate(anno_list))
elif len(ch) == 2:
anno_blist = bundle_list(_anno_list, 2)
else:
raise ValueError("Set `anno_ch` in ('DNA', 'actin') or all")
return dict((k, v) for k, v in enumerate(anno_blist))