#!/usr/bin/env python3
"""
Use ``img_DoG`` from UNRAVEL to apply Difference of Gaussians to a single image.
Difference of Gaussians:
- Sigma1 and sigma2 are the standard deviations for the first and second Gaussian blurs
- Simga2 (the larger blur) should be ~ 1.0 to 1.5 times the radius of these features of interest
- E.g., if nuclei have a radius of ~1.5 to 2.5 pixels, sigma2 might be 1.5 to 3.0
- Sigma1 (the smaller blur) should be smaller than the size of the features you want to keep, ideally around the size of the noise
- E.g., if noise is ~1 pixel in size, sigma1 might be 0.5 to 1
- The ratio of simga2 to sigma1 should ideally be at least 1.5 to 2. This helps ensure that the blurring difference is significant enough to highlight the features of interest.
Note:
- This command is intended to test the DoG method on a single image.
- 2D DoG is not implemented in vstats_prep.
- DoG could be added to vstats_prep in the future if needed.
- 3D spatial averaging and 2D rolling ball background subtraction are used in vstats_prep instead.
Usage:
------
img_DoG -i input.tif -g1 1.0 -g2 2.0 [-o output.tif] [-v]
"""
import cv2
import numpy as np
from rich import print
from rich.traceback import install
from unravel.core.help_formatter import RichArgumentParser, SuppressMetavar, SM
from unravel.core.config import Configuration
from unravel.core.utils import log_command, verbose_start_msg, verbose_end_msg
[docs]
def parse_args():
parser = RichArgumentParser(formatter_class=SuppressMetavar, add_help=False, docstring=__doc__)
reqs = parser.add_argument_group('Required arguments')
reqs.add_argument('-i', '--input', help='Path to the input TIFF file.', required=True, action=SM)
reqs.add_argument('-g1', '--sigma1', help='Sigma for the first Gaussian blur in DoG (targets noise)', default=None, required=True, type=float)
reqs.add_argument('-g2', '--sigma2', help='Sigma for the second Gaussian blur in DoG (targets signal).', default=None, required=True, type=float)
opts = parser.add_argument_group('Optional arguments')
opts.add_argument('-o', '--output', help='Path to save the output TIFF file.', default=None, action=SM)
general = parser.add_argument_group('General arguments')
general.add_argument('-v', '--verbose', help='Increase verbosity. Default: False', action='store_true', default=False)
return parser.parse_args()
# TODO: Add support for other image types and 3D images.
[docs]
def load_tif(tif_path):
'''Load a single tif file using OpenCV and return ndarray.'''
img = cv2.imread(tif_path, cv2.IMREAD_UNCHANGED)
if img is None:
raise FileNotFoundError(f'Could not load the TIFF file from {tif_path}')
return img
[docs]
def difference_of_gaussians(img, sigma1, sigma2):
'''Subtract one blurred version of the image from another to highlight edges.'''
blur1 = cv2.GaussianBlur(img, (0, 0), sigma1)
blur2 = cv2.GaussianBlur(img, (0, 0), sigma2)
dog_img = cv2.subtract(blur1, blur2)
return dog_img
[docs]
def save_tif(img, output_path):
'''Save an image as a tif file.'''
cv2.imwrite(output_path, img)
[docs]
@log_command
def main():
install()
args = parse_args()
Configuration.verbose = args.verbose
verbose_start_msg()
# Load the image
img = load_tif(args.input)
# Apply difference of Gaussians if sigmas are provided
img = difference_of_gaussians(img, args.sigma1, args.sigma2)
print(f'Applied Difference of Gaussians with sigmas {args.sigma1} and {args.sigma2}.')
# Save the processed image
output_path = args.output if args.output is not None else args.input.replace('.tif', f'_DoG{args.sigma2}-{args.sigma1}.tif')
save_tif(img, output_path)
verbose_end_msg()
if __name__ == '__main__':
main()