dicom2jpg

Converts DICOM to JPG/PNG/BMP/TIFF and numpy.ndarray

Installation

pip install dicom2jpg

Introdunction

import dicom2jpg

dicom_img_01 = "/Users/user/Desktop/img01.dcm"
dicom_dir = "/Users/user/Desktop/Patient_01"
export_location = "/Users/user/Desktop/BMP_files"

# convert single DICOM file to jpg format
dicom2jpg.dicom2jpg(dicom_img_01)  

# convert all DICOM files in dicom_dir folder to png format
dicom2jpg.dicom2png(dicom_dir)  

# convert all DICOM files in dicom_dir folder to bmp, to a specified location
dicom2jpg.dicom2bmp(dicom_dir, target_root=export_location) 

# convert single DICOM file to numpy.ndarray for further use
img_data = dicom2jpg.dicom2img(dicom_img_01)

# convert DICOM ByteIO to numpy.ndarray
img_data = dicom2jpg.io2img(dicomIO)

dicom2jpg converts DICOM images to JPG/PNG/BMP/TIFF formats and to numpy.ndarray. It piplines the lookup transformations by applying Modality LUT, VOI LUT, and Presentation LUT to the images, which makes output files looks like what we see on standard DICOM viewers.

dicom2jpg.dicom2jpg(origin, target_root=None, anonymous=False, multiprocessing=True)

dicom2jpg.dicom2png(origin, target_root=None, anonymous=False, multiprocessing=True)

dicom2jpg.dicom2bmp(origin, target_root=None, anonymous=False, multiprocessing=True)

dicom2jpg.dicom2tiff(origin, target_root=None, anonymous=False, multiprocessing=True)

• origin can be a single DICOM file, a folder, or a list/tuple of file/folder

• target_root would be the root folder of the first file/folder if not specified

• exported files paths would be

  target_root / Today / PatientID_Filetype / StudyDate_StudyTime_Modality_AccNum / Ser_Img.Filetype

• anonymous files paths would be

  target_root / Today / Patient_SerialNum / ModalitySerialNum_Modality / Ser_Img.Filetype

dicom2jpg.dicom2img(origin)

dicom2jpg.io2img(dicomIO)

• converting dicom files or ByteIO to ndarray
• ndarray is in 8 bit; RGB format if it's a color image

Image examples

CT	MR	CXR

Todo

• Support multi-frame images
• Image compression
• Support overlays

Performance

• Environment: Windows10, Jupyter Notebook, Python 3.8.10
• 598MB 1873 files {'CT': 1528, 'CR': 52, 'MR': 174, 'DX': 36}
• Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz. 4 Cores (hyper-threading off)
• Tested on Ramdisk (no physical HDD was tortured :P)

multiprocessing	anonymous	duration (seconds)
False	True	154.6-159.7
True	True	79.2-82.9
False	False	157.9-162.8
True	False	56-58.5