MajwareMAJWARE
Video tutorialDICOM & Medical Imaging

DICOM Explained in 3 Minutes: From Pixels to PACS

Duration 3:04 · Published 11 November 2026 · on YouTube (Majware | Digital Healthcare, opens in a new tab)

All images in this video are synthetic. No patient data is used. The player loads from YouTube, in privacy-enhanced mode, when you press play.

Discover how DICOM makes medical imaging possible. In this beginner-friendly video, we introduce DICOM, explain how medical images combine pixel data with essential metadata, and explore how imaging systems and PACS work together to store, exchange, and display medical images. A quick introduction to the technology behind modern digital radiology.

Who it's for: PACS administrators, radiology technologists, healthcare IT engineers, and medical imaging students.

What You'll Learn

As listed in the video's description on YouTube.

  • How pixels are stored: Rows/Columns, Bits Allocated vs Bits Stored, Pixel Representation, Photometric Interpretation, Pixel Spacing
  • How metadata works: tags, Value Representations (VR), and Explicit vs Implicit VR
  • The anatomy of a Part 10 file: 128-byte preamble, "DICM" prefix, File Meta Information, Data Set, Pixel Data
  • How a CT image is displayed: stored value → Rescale Slope/Intercept → Hounsfield Units → Window/Level
  • The Patient → Study → Series → SOP Instance hierarchy, and multi-frame objects
  • DICOM networking: C-STORE, C-FIND, C-MOVE, C-GET, and DICOMweb (QIDO-RS, WADO-RS, STOW-RS)

Chapters

Chapter times from the video's description on YouTube. Choose one to play from that point.

  1. Play from 0:00What Is DICOM?
  2. Play from 0:18Understanding Image Pixels
  3. Play from 0:44DICOM Metadata and Tags
  4. Play from 1:08DICOM File Anatomy
  5. Play from 1:32From Pixel Data to a CT Image
  6. Play from 1:59Patient, Study, Series and Images
  7. Play from 2:22DICOM Communication and PACS
  8. Play from 2:46Summary

Background · written for this page, not a transcript

DICOM in brief

DICOM (Digital Imaging and Communications in Medicine) is the international standard for medical images and the information that travels with them. It defines both a file format and a network protocol, so that a CT scanner, a PACS and a viewer from different manufacturers can exchange images and read them the same way. The standard is maintained by the DICOM Standards Committee, with NEMA as its secretariat, and is also published as ISO 12052. The current edition (opens in a new tab) is free to read online.

Pixel data is a grid of numbers

The picture itself is stored in the Pixel Data attribute as a grid of numbers. Rows and Columns give the size of the grid. Each value takes up Bits Allocated bits in the file, of which Bits Stored carry the value: a CT image commonly allocates 16 bits and stores 12.

Three more attributes say how to read those numbers. Pixel Representation says whether they are unsigned (0) or signed (1). Photometric Interpretation says what they mean on screen: in MONOCHROME2 the lowest value is shown as black, in MONOCHROME1 as white, and RGB holds colour. Pixel Spacing gives the physical distance between pixel centres in millimetres, which is what makes measurements possible.

Metadata: tags and value representations

Everything else about the image is metadata, stored as data elements. Each element has a tag, a pair of hexadecimal numbers written as (group,element), such as (0010,0020) for Patient ID. It also has a value representation (VR), the data type of its value: PN for a person's name, DA for a date, UI for a unique identifier, US for an unsigned 16-bit integer.

With explicit VR, the two-letter VR is written in the file next to each tag. With implicit VR it is left out, and the reader has to look it up in the DICOM data dictionary. Which of the two a file uses is set by its Transfer Syntax UID, along with byte order and compression. The transfer syntax troubleshooting guide covers what happens when two systems disagree about it.

Inside a DICOM file

A DICOM file, as defined in Part 10 of the standard, starts with a 128-byte preamble followed by the four characters DICM. Next comes the File Meta Information (group 0002), always encoded as explicit VR little endian, which records among other things the transfer syntax of the rest of the file. After it comes the data set: patient, study, series and image attributes, usually ending with the pixel data.

From stored values to a CT image

Stored pixel values are not yet in clinical units. For CT, a viewer first applies Rescale Slope and Rescale Intercept: stored value × slope + intercept gives Hounsfield units (HU), on a scale where water is 0 and air is about −1000. A slope of 1 with an intercept of −1024 is a common pairing.

A screen shows far fewer grey levels than CT can store, so the viewer then picks a range of HU to spread across them, set by Window Center and Window Width, often called window/level. A soft-tissue window might be centred at 40 HU with a width of 400; a lung window at −600 HU with a width of 1500.

Patient, study, series, instance

DICOM files every object in a four-level hierarchy. A patient has studies, each one imaging examination identified by its Study Instance UID. A study holds series, sets of images acquired or reconstructed together on one modality, each with a Series Instance UID. Each object in a series is an instance with its own SOP Instance UID. An instance usually holds one image, but a multi-frame object holds many, counted in Number of Frames.

Where PACS comes in

A PACS (picture archiving and communication system) receives images from modalities, stores them, and serves them to viewing workstations. The classic DICOM network services do that work: C-STORE sends objects to the archive, C-FIND queries what it holds at patient, study, series or image level, and C-MOVE and C-GET retrieve them. C-MOVE asks the archive to send the objects to a named destination, while C-GET returns them on the same connection.

DICOMweb provides the same over HTTP: QIDO-RS to query, WADO-RS to retrieve and STOW-RS to store. For association negotiation, AE titles and the rest of the networking detail, see the DICOM Protocol reference.

DICOM Attributes Behind These Topics

Each links to its page in the DICOM tag reference, with its value representation, where it is used and the problems it commonly causes.

More tutorials will appear here as they are published.