How do we quickly learn the meaning of MATLAB's dicominfo metadata?

I have been unable to find MathWorks documentation explaining each of the dicominfo metadata entries, e.g. the meaning of [0;1] for "FileMetaInformationVersion". Must I simply look through the DICOM Standard? Do you have other recommendations? (My goal is to align contours with the dose distribution.)

Indeed, whereas MATLAB reports an ROIContourSequence, the Standard's section 'C.8.8.6 ROI Contour Module' does explain its structure fields 'ROIDisplayColor', 'ContourSequence', 'ReferencedROINumber'.

The DICOM Standard is so verbose, however, that I think it would save time if MathWorks would add a lookup function to provide a brief explanation (perhaps even quoting from the Standard) for each metadata entry. For example,

dicommeaning(info.ROIContourSequence.ReferencedROINumber)

ans = Uniquely identifies the referenced ROI described in the Structure Set ROI Sequence (3006,0020). (C.8.8.6 ROI Contour Module of the DICOM Standard)

It seems to me this is something WolframAlpha would do. Would implementing this feature put MathWorks in undesired competition?

Or is there simply not enough demand or need for it? Perhaps a simple Internet search accomplishes the same goal. For example, this OTpedia article results from searching Bing: "What is SOP Instance UID?" Similarly, searching Bing: "File Meta Information Version DICOM" yields Dabsoft's explanation (apparently an HTML version of the DICOM Standard's part10.pdf file).

In your experience, are Internet searches the fastest way to learn what is returned by dicominfo, rather than sifting through the DICOM Standard?

Answer 1

[Daniel Bridges]

The fastest way to learn the meaning of dicominfo's DICOM metadata in my experience is to read the DICOM Standard using text-searchable PDFs; Internet searches can also yield the same results.

The DICOM Standard appeared daunting at first, but once I became oriented it was straightforward.

Answer 2

[Walter Roberson]

The DICOM Standard is somewhat large, and is constantly changing. It would be a lot of work for Mathworks to add a descriptive dictionary of the terms.

There are multiple books about DICOM; see http://dicom.nema.org/documents/books.pdf and https://www.quora.com/What-are-the-best-online-book-references-for-DICOM-related-algorithms

[Walter Roberson]

Wolfram Alpha aspires to be a knowledge base of all factual knowledge.

"Wolfram Alpha is written in 15 million lines of Wolfram Language code[12] and runs on more than 10,000 CPUs" (wikipedia)

It isn't a matter of undesired competition: it is a matter of resources, and market plans.

I think it is fairly safe to say that Mathworks has never especially invested in DICOM. They added a routine to read the files; they probably used a third-party library rather than writing it themselves.

There are companies that are strong players in DICOM, including strong analysis and visualization and even tools for repair of images, and which implement multiple parts of the DICOM standard (DICOM is a lot more than just an image format.) Hospital data management systems based around DICOM are big business (and very expensive to produce to ISO standards, and passing FDA clearance.) It is expensive to do well in that market.

I would not imagine that Mathworks is headed in that direction. Look at where they have been producing new features: a lot of it is in the Arduino / Raspberry Pi / FPGA markets of code generation, of simulation and modelling.

Mathworks does not even especially publicize SPM, the third-party (free) package that is widely used for MRI.

If you were a Mathworks Product Manager, would you be confident in predicting that the time spent preparing and maintaining a data tool (not so much code) about DICOM would be likely to draw in enough extra business to justify the cost and effort?

[Daniel Bridges]

Thanks for additional context. I was hopeful for more freely available resources, or content available from MathWorks I hadn't discovered. Some of the links you've shared don't work, but I understand you're simply recommending books in general.

The question then becomes which book best explains the DICOM metadata ... Incidentally, the Philips download URL appears to have changed to an internal-use HTTP server instead of the FTP server referenced by NEMA.

Perhaps I can manage by referring to the NEMA's DICOM PS3.1 2016b (what I've been calling "the DICOM Standard") together with Internet searches...

Edit: Looks like you were right to say it's constantly changing: There's already DICOM PS3.1 2016c.

[Walter Roberson]

The TIFF tag documentation talks about DICOM and how some tags were being used by agreement for a 2 year period and then other tags were going to come into effect... When I saw that it became apparent to me that DICOM was changing yearly.