Decode DICOM files for medical imaging

Stay organized with collections Save and categorize content based on your preferences.

View on TensorFlow.org Run in Google Colab View source on GitHub Download notebook

Overview

This tutorial shows how to use tfio.image.decode_dicom_image in TensorFlow IO to decode DICOM files with TensorFlow.

Setup and Usage

Download DICOM image

The DICOM image used in this tutorial is from the NIH Chest X-ray dataset.

The NIH Chest X-ray dataset consists of 100,000 de-identified images of chest x-rays in PNG format, provided by NIH Clinical Center and could be downloaded through this link.

Google Cloud also provides a DICOM version of the images, available in Cloud Storage.

In this tutorial, you will download a sample file of the dataset from the GitHub repo

  • Xiaosong Wang, Yifan Peng, Le Lu, Zhiyong Lu, Mohammadhadi Bagheri, Ronald Summers, ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases, IEEE CVPR, pp. 3462-3471, 2017
curl -OL https://github.com/tensorflow/io/raw/master/docs/tutorials/dicom/dicom_00000001_000.dcm
ls -l dicom_00000001_000.dcm
% Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   164    0   164    0     0    600      0 --:--:-- --:--:-- --:--:--   598
100 1024k  100 1024k    0     0  1915k      0 --:--:-- --:--:-- --:--:-- 1915k
-rw-rw-r-- 1 kbuilder kokoro 1049332 Nov 22 03:47 dicom_00000001_000.dcm

Install required Packages, and restart runtime

try:
  # Use the Colab's preinstalled TensorFlow 2.x
  %tensorflow_version 2.x 
except:
  pass
pip install tensorflow-io

Decode DICOM image

import matplotlib.pyplot as plt
import numpy as np

import tensorflow as tf
import tensorflow_io as tfio

image_bytes = tf.io.read_file('dicom_00000001_000.dcm')

image = tfio.image.decode_dicom_image(image_bytes, dtype=tf.uint16)

skipped = tfio.image.decode_dicom_image(image_bytes, on_error='skip', dtype=tf.uint8)

lossy_image = tfio.image.decode_dicom_image(image_bytes, scale='auto', on_error='lossy', dtype=tf.uint8)


fig, axes = plt.subplots(1,2, figsize=(10,10))
axes[0].imshow(np.squeeze(image.numpy()), cmap='gray')
axes[0].set_title('image')
axes[1].imshow(np.squeeze(lossy_image.numpy()), cmap='gray')
axes[1].set_title('lossy image');
2021-11-22 03:47:53.016507: E tensorflow/stream_executor/cuda/cuda_driver.cc:271] failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected

png

Decode DICOM Metadata and working with Tags

decode_dicom_data decodes tag information. dicom_tags contains useful information as the patient's age and sex, so you can use DICOM tags such as dicom_tags.PatientsAge and dicom_tags.PatientsSex. tensorflow_io borrow the same tag notation from the pydicom dicom package.

tag_id = tfio.image.dicom_tags.PatientsAge
tag_value = tfio.image.decode_dicom_data(image_bytes,tag_id)
print(tag_value)
tf.Tensor(b'58', shape=(), dtype=string)
print(f"PatientsAge : {tag_value.numpy().decode('UTF-8')}")
PatientsAge : 58
tag_id = tfio.image.dicom_tags.PatientsSex
tag_value = tfio.image.decode_dicom_data(image_bytes,tag_id)
print(f"PatientsSex : {tag_value.numpy().decode('UTF-8')}")
PatientsSex : M