LightDICOM is a fast as light, light as feather Python3 package for managing DICOM files.
The only 3rd party dependency is numpy, thus fast and light.
- Source Code: Github
- Korean version: 한국어 README
- Bug report/Feature request: Open a Github Issue or contact me at
ryoungwoo.jang@vuno.co - Benchmarks: WIP
- DEMO: WIP
- Documentation: WIP
from LightClass import LightDCMClass
lc = LightDCMClass()
lc.path = path_to_dicom
d = lc.get_data('0010,0010') # Read tag (0010, 0010), which is "Patient's Name"
npy = lc.read_pixel() # Read pixel values
Or
from LightClass import LightDCMClass
lc = LightDCMClass(path=path_to_dicom)
d = lc.get_data('0010,0010') # Read tag (0010, 0010), which is "Patient's Name"
npy = lc.read_pixel() # Read pixel values
lc = LightDCMClass()
lc.path = path_to_dicom
# Equivalent code : lc = LightDCMClass(path=path_to_dicom)
all_headers = lc.read_all(with_pixel=True)
In .read_all method, there are two arguments:
- with_pixel : If True, read header information with pixel values. Else, .read_all will not read pixel values. There will be no significant time difference, yet memory efficiency will be different. False uses less memory. That's the difference.
- resize_pixel : If True, pixel value will be reshaped according to width (0028,0010) and height (0028,0011) information. I guess that in most situations, setting .resize_pixel to False will not be required.
- Python 3.x
- Numpy
If numpy is not installed, first install it by pip, conda or other method of your choice.
- install from source
git clone https://github.com/jryoungw/lightdicom - pypi, conda, etc. are not yet supported. (WIP)
This is a prototype version, so do remember that bugs or unexpected behaviors may exist.
Whenever bugs are found, please contact via the aforementioned methods. PRs are welcome!