Inferensys

Glossary

DICOM File Meta Information

The mandatory header at the beginning of a DICOM Part 10 file that contains the File Preamble, DICOM prefix, and critical elements like the Transfer Syntax UID and Media Storage SOP Class UID.
Stylish WeWork-like workspace with hot desks and document wall, professional searching through enterprise knowledge base on a mounted ultrawide display, warm industrial pendants overhead.
DICOM PART 10 HEADER

What is DICOM File Meta Information?

The mandatory header at the beginning of every DICOM Part 10 file that provides the essential context required for any application to correctly parse the subsequent data set.

DICOM File Meta Information is the mandatory header at the beginning of a DICOM Part 10 file, containing the 128-byte File Preamble, the 4-byte DICM prefix, and a set of encoded DICOM Tags that describe the file's identity and encoding. It functions as a self-describing wrapper, providing the critical Transfer Syntax UID and Media Storage SOP Class UID that a parser must read before interpreting the main data set.

This header is always encoded using Explicit VR Little Endian transfer syntax to guarantee universal readability before the actual compression scheme is negotiated. Key elements include the (0002,0010) Transfer Syntax UID, which specifies the byte ordering and compression used for the rest of the file, and the (0002,0002) Media Storage SOP Class UID, which identifies the type of object, such as a CT Image or Structured Report, enabling a reader to instantiate the correct parsing logic.

DICOM PART 10 HEADER

Key Elements in the File Meta Information (Group 0002)

The mandatory header at the beginning of every DICOM Part 10 file. It contains the critical identification and encoding parameters required for any system to parse the subsequent data set.

01

Media Storage SOP Class UID (0002,0002)

Uniquely identifies the type of data object stored in the file, such as a CT Image or MR Spectroscopy. This is the most critical interoperability key, defining the exact Information Object Definition (IOD) and SOP Class to which the data set conforms.

  • Example Value: 1.2.840.10008.5.1.4.1.1.2 (CT Image Storage)
  • Role: Dictates which DICOM tags are mandatory and how the pixel data should be interpreted.
UID
Data Type
1
Cardinality
02

Transfer Syntax UID (0002,0010)

Defines the encoding rules used to serialize the entire data set into a byte stream. This tag specifies the byte ordering (Little Endian vs. Big Endian), whether encapsulation is used, and the compression algorithm applied to the pixel data.

  • Example Value: 1.2.840.10008.1.2.4.70 (JPEG Lossless, First-Order Prediction)
  • Role: Without this, a parser cannot correctly read multi-byte values or decompress the image.
UID
Data Type
1
Cardinality
03

Media Storage SOP Instance UID (0002,0003)

The globally unique identifier for this specific file instance. Combined with the Media Storage SOP Class UID, this pair forms a unique key that distinguishes this exact scan from every other scan in the world.

  • Example Value: 1.2.840.113619.2.55.3.2831181172.699.1728884896.511.2
  • Role: Essential for database indexing, retrieval, and preventing duplicate storage in a PACS or VNA.
UID
Data Type
1
Cardinality
04

Implementation Class UID (0002,0012)

Uniquely identifies the software implementation that wrote this file. This allows receiving systems to identify the source of the data and apply vendor-specific workarounds if necessary.

  • Example Value: 1.2.276.0.7230010.3.0.3.6.7 (a specific toolkit version)
  • Role: Critical for debugging interoperability failures and tracking the provenance of a file during DICOM Conformance Statement analysis.
UID
Data Type
1
Cardinality
05

Implementation Version Name (0002,0013)

A human-readable string identifying the software version that created the file, used for logging and debugging purposes. It complements the machine-readable Implementation Class UID.

  • Example Value: OFFIS_DCMTK_368
  • Role: Allows support engineers to quickly identify the generating software stack without decoding a UID.
SH
Value Representation
3
Type
06

File Preamble & DICOM Prefix

The first 132 bytes of the file, consisting of a 128-byte File Preamble (often zeroed or used for application-specific notes) followed by a 4-byte DICOM Prefix (D, I, C, M).

  • Role: This magic number distinguishes a DICOM file from other file formats and signals the start of the File Meta Information header.
128+4
Byte Length
"DICM"
Magic Number
DICOM FILE META INFORMATION

Frequently Asked Questions

Essential questions about the mandatory header that enables DICOM Part 10 file parsing, including the File Preamble, DICOM prefix, and critical identification elements like Transfer Syntax UID and Media Storage SOP Class UID.

DICOM File Meta Information is the mandatory header at the beginning of every DICOM Part 10 file that provides the essential identification and encoding context required to parse the subsequent data set. Without this header, a parser cannot determine the byte ordering, compression scheme, or the type of SOP Instance contained in the file. The File Meta Information is encapsulated as a DICOM data set with a Group 0002 tag prefix, meaning all its elements belong to group (0002,xxxx). It begins after a 128-byte File Preamble and a 4-byte DICOM prefix (DICM), which together serve as a compatibility buffer and format verification marker. The most critical elements within this header are the Transfer Syntax UID (0002,0010), which defines the encoding rules, and the Media Storage SOP Class UID (0002,0002), which identifies the type of object stored. The File Meta Information must be encoded using Explicit VR Little Endian transfer syntax regardless of the encoding used for the rest of the file, ensuring that any conforming parser can always read the header to discover how to interpret the remaining bytes.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.