DICOM Support in Image Processing Toolbox
Digital Imaging and Communications in Medicine (DICOM) is a highly standardized imaging format used to store and transmit medical imaging files across devices and networks. The DICOM format combines image data with metadata that describes the patient, the imaging procedure, and the spatial referencing information. The structure, storage, and transmission of DICOM files is governed by the DICOM standard, available on the official DICOM website. The standard defines separate Information Object Definitions (IODs) for modalities and applications such as computed tomography (CT), magnetic resonance imaging (MRI), and radiotherapy (RT).
MATLAB® provides support for reading and writing DICOM files, as well as working with DICOM image data and metadata. You can extract and process image data using toolbox functions, and you can search and update the metadata attributes. MATLAB is compatible with most DICOM IODs, and can write new DICOM files for certain IODs that fully conform to the DICOM standard.
MATLAB supports working with DICOM files. There is no support for working with DICOM network capabilities.
Read and Display DICOM Image Data
Explore directories with multiple DICOM series using the DICOM Browser app or
function. Read 2-D image data from a DICOM series by using the
dicomread function or 3-D image data by using the
function. For more information, see Read Image Data from DICOM Files. You can view imported DICOM images using toolbox
display functions such as
You can process the image data you read from DICOM files using operations such as image filtering, registration, segmentation, and labeling. For an example that shows how to segment and calculate region properties in medical image data, see Segment Lungs from 3-D Chest Scan.
Work with DICOM Metadata
Import DICOM metadata using the
dicominfo function, which creates a MATLAB structure specifying the name and value of each metadata attribute in the
file. For more information, see Read Metadata from DICOM Files.
List all attributes of a metadata structure in the Command Window by using the
dicomdisp function, or search for specific attributes by name using the
function. Update specific attribute values using the
function, or remove all personally identifying information from a DICOM metadata structure
dicomanon function. For an example that shows
how to anonymize a DICOM file, see Remove Confidential Information from DICOM File .
When processing DICOM files, MATLAB uses a data dictionary file that defines standard DICOM metadata attributes.
You can view or update the current data dictionary file using the
dicomdict function, or search the data dictionary for specific attributes
Write New DICOM Files
Write images and metadata to new DICOM files using the
dicomwrite function. The toolbox writes the computed tomography, magnetic
resonance, and secondary capture (a modality-independent object definition) IODs with
validation, which ensures that the new file contains all metadata attributes required by the
DICOM standard. For detailed information, see Write Image Data to DICOM Files and
Create New DICOM Series.
Work with DICOM-RT Contour Data
The DICOM-RT Structure Set is an IOD specific to radiotherapy applications. The DICOM-RT
metadata includes contour data for ROIs, such as tumors and organs, used in radiation
treatment planning. You can extract ROI contour data to create a
Plot contours, add or delete contours, and create a new DICOM-RT metadata structure
object functions. For an example, see Add and Modify ROIs of DICOM-RT Contour Data.
object function to convert contour data into a binary mask, such as to view ROIs overlaid on
image data or to label image pixels. For an example, see Create and Display 3-D Mask of DICOM-RT Contour Data.
Prepare DICOM Files for Deep Learning Workflows
You can use medical image data to train deep learning networks to perform tasks such as
image denoising, segmentation, and registration. You can use
pixelLabelDatastore (Computer Vision Toolbox) objects that contain DICOM files to train a deep learning
network. For details, see Create Image Datastore Containing DICOM Images and Create Image Datastore Containing Single and Multi-File DICOM Volumes. For more information
about how to use image datastores to train deep learning networks, see Preprocess Images for Deep Learning.
These examples show applications of deep learning in medical image analysis.