Python读取与显示DICOM图像入门：所需库及代码示例

Hey there! I’ve walked a lot of newbies through DICOM processing in Python, so let’s break this down into simple, practical steps that make sense for someone just getting started.

Essential Libraries for DICOM Work #

First, you’ll need a couple of core packages—these are the standard tools used by both beginners and experienced developers:

• pydicom: The main library for reading and parsing DICOM files. It handles both the metadata (like patient info, scan settings) and the actual pixel data of the image.
• matplotlib: A straightforward plotting library that lets you display the DICOM image once you’ve extracted the pixel data.
• (Optional) SimpleITK: If you plan to dive into advanced tasks like image segmentation or registration later, this is a powerful tool—but start with the first two for the basics.

Install them with these terminal commands:

pip install pydicom matplotlib

Step-by-Step Code to Read and Display a DICOM Image #

Here’s a fully commented example that walks you through every part. I’ll explain each step so you understand what’s going on:

# Import the libraries we need
import pydicom
import matplotlib.pyplot as plt

# 1. Load your DICOM file
# Replace 'your_dicom_file.dcm' with the actual path to your file
dicom_dataset = pydicom.dcmread("your_dicom_file.dcm")

# Optional: Peek at the metadata (patient name, scan dimensions, etc.)
# Uncomment the line below to see all the stored metadata
# print(dicom_dataset)

# 2. Extract the pixel data from the DICOM object
pixel_data = dicom_dataset.pixel_array

# 3. Normalize pixel values for proper display
# Most DICOMs use 16-bit values (0-65535), but matplotlib expects 8-bit (0-255)
# This scales the values so the image looks clear instead of washed out or too dark
normalized_data = (pixel_data - pixel_data.min()) / (pixel_data.max() - pixel_data.min()) * 255
normalized_data = normalized_data.astype("uint8")

# 4. Display the image
plt.figure(figsize=(10, 10))
plt.imshow(normalized_data, cmap="gray")  # Grayscale is standard for medical images
plt.title("DICOM Image")
plt.axis("off")  # Hide axis ticks for a cleaner view
plt.show()

Quick Tips for Beginners #

• Why normalize? Skipping this step often leads to images that look wrong—DICOMs store pixel values in a wide range, and we need to scale them down to what display tools can handle.
• Multi-frame DICOMs: The code above works for single-frame images (like X-rays or individual CT slices). If you have a series of scans (a 3D volume), pixel_data will be a 3D array—you’ll need to loop through the frames to display each one.
• Accessing metadata: You can pull specific fields directly, like dicom_dataset.PatientName for the patient’s name or dicom_dataset.Rows for the image height.

内容的提问来源于stack exchange，提问作者s.sowmiya