Extract images from Dicoms of Heidelberg OCT machine.

Heidelberg_dcm_to_png.py

"""
Image extractor for Heidelberg OCT B-SCANS.

Author: Jian Shi
Date created: 28/08/2020
"""

from multiprocessing import Pool
import matplotlib.pyplot as plt
import pydicom
import os
import logging

__author__ = "Jian Shi"
__license__ = "MIT"
__version__ = "0.1"
__maintainer__ = "Jian Shi"
__email__ = "[email protected]"

logging.root.setLevel(logging.NOTSET)


def dcm_to_png(fp):
    fn = fp.replace('.dcm', '.png')
    # 16 bits to 8 bits
    img = pydicom.read_file(fp).pixel_array ** (1 / 2)
    plt.imsave(fn, 255 - img, cmap='gray')


if __name__ == '__main__':
    import argparse
    parser = argparse.ArgumentParser(description='Extract images from Dicoms of Heidelberg OCT machine.')
    parser.add_argument('--folder', required=True, help='Target folder path.')
    parser.add_argument('--process', metavar='N', type=int, default=20,
                        help='Number of processes can be used.')
    args = parser.parse_args()

    process = args.process
    pt = args.folder

    p = Pool(process)
    fp_list = []
    logging.info("Reading dcms.")
    for root, _, files in os.walk(pt):
        for f in files:
            if f.endswith('.dcm'):
                fp_list.append(os.path.join(root, f))
    logging.info(f"Get {len(fp_list)} files. Start processing with {process} processes.")
    res = p.imap(dcm_to_png, fp_list)
    for i, x in enumerate(res):
        progress = i / len(fp_list) * 100
        print(f'Finished {progress:.2f}%.', end='\r')
    logging.info("Finished.")