Hi Dr. Kindlmann,
Thank you for this detailed reply.
No I wasn't expecting it to be a GUI app. I am the software designer
here at a small geophysics company, so software library and command line
is fine. If I find it looks useful for either job (time series graph,
or magnetic field maps) I can code around it to do what I want and
present what I want.
I guess my comment on the starting web page
(http://teem.sourceforge.net/) would be that it took too many clicks to
find any tangible examples or demos (like
http://teem.sourceforge.net/nrrd/demo.html), and (as I see now) they
were only under the specific library links (nrrd, gage, mite, ten as
linked from that start page), so I initially glossed over the demo-y
links because I figured I didn't want to drill down that specifically to
individual libraries right away.
I have now read through a lot of the demo pages and the paper on
curvature based transfer functions so it makes a lot more sense.
My scenario is airborne geophysical surveys (magnetic and radiometric).
We're flying sensors around on helicopters in a grid pattern over
regions on the order of tens of kilometers, with line spacings usually
50 to 200 m. So although the scale is vastly different from medical,
the tasks seem somewhat similar. Also this is not a 3d grid of data
points like an MRI, only 2d. There is also the issue of the idealized
flight path (perfectly straight at constant altitude) which we intend to
send our sensors through, versus the actual flight lines affected by the
vagaries of topography, wind, weather, pilot skill, etc. So we have
many filters and corrections from many sources and methods, which it
strikes me must be similar to tasks done in medical imaging.
Thanks again for your time in replying, and for providing this
impressive looking tool.
On 26/07/13 05:22 PM, Gordon L. Kindlmann wrote:
> Hello Martin,
> Thanks for the query and the opportunity to describe Teem.
> One of the first things to know about Teem is that it is a collection of libraries, and a few command-line tools, but there is no GUI, or anything with a network connection. Nor is there much stuff that is specific to any one field of study, with the notable exception of the diffusion-weighted MRI stuff in the "ten", and "elf" libraries. So without doing any research, my guess about the answer to "is Teem used for geophysics" is "probably not, but you could try coding something up :) ".
> Teem is probably more low-level that you think - being a set of libraries, it is used by people who write applications that link against libraries, not by people looking for applications. You can write applications that are useful for various domains, and draw on functions in Teem, but there's nothing "out of the box" that suits your stated needs (web-accessible graphing of time-series data and mapping magnetic fields).
> This may be an unsatisfying answer, and it doesn't answer the "what is it good for" question. I would say that in Teem, just to highlight a few of the libraries: "nrrd" is good for raster data manipulation (see the unu commands), "gage" is good for creating the abstraction of continuous fields from raster data via convolution, "ten" is good for doing things with tensor fields (see the tend commands), and "pull" is good for particle-based feature extraction in continuous fields. Except for possibly "ten", none of these are very application specific, because they weren't coded or designed for specific applications. They are just libraries.
> The mailing list traffic is low because most people who pick up Teem figure out how to use it by experimentation or looking at the source code. Also, Teem has a relatively small user community (compared to things like VTK or ITK). But don't mistake this for being moribund:
> this documents some general properties of the project. It says there haven't been recent commits because Sourceforge recently switched to new SVN servers. Prompted by your email, I just updated the ohloh page to include the new SVN server.
> The last release was in December:
> and that page lists what is new relative to the previous release. A patch release is coming out next month.
> Answering "how many people use Teem" is tricky. Anyone can get the source and use it without telling us, so there's no central accounting of number of users.
> Good luck,
> On Jul 25, 2013, at 4:36 PM, martin.muc@... wrote:
>> Hey list,
>> I've stumbled on teem because of a coincidence in your file format
>> name, nrrd, matching a search for "rrd". I was actually looking for
>> Debian packages similar to rrdtool (round robin database - for storing
>> and graphing time-series data).
>> Actually I have two tasks, web-accessible graphing of time-series data
>> that is more frequent than 1 Hz (which is a limitation of rrdtool),
>> and mapping magnetic fields (or, technically, magnetic intensity).
>> And reading about teem has gotten me intrigued about teem's potential
>> suitability for both of these tasks.
>> I've clicked around and read on teem.sourceforge.net. Clicked on
>> "air"... fairly terse... okay it's got some macros. I clicked on
>> "ell"... okay it's for linear algebra... fairly terse. I've clicked a
>> smattering of teem-users@ mailing list threads in the archive,
>> randomly from 2004, 2008, 2012, mostly discussing compile problems,
>> bug fixes, etc... Screen shots on
>> http://sourceforge.net/projects/teem/ are text output of processing
>> So, aside from "Teem is a coordinated group of libraries for
>> representing, processing, and visualizing scientific raster data",
>> what _IS_ teem? What is it being used for? Aside from mentioning
>> "Healthcare Industry" in the intended audience, and a few of the
>> thumbnail images (not the Mr. T one) for the different libraries being
>> slightly suggestive of medical X-rays or MRIs, I can't tell what teem
>> really does or is good at, It is all so general (as science and math
>> of course are), I can't tell what teem _is_ or what it _does_.
>> Okay, never mind. I've just found
>> http://teem.sourceforge.net/nrrd/demo.html I will read some of these
>> examples and I guess if it looks promising, will download and try teem
>> with some of these tutorials.
>> Also, is anyone using teem for geophysics? Broadly, parts of our
>> geophysics output (mapping) _are_ "scientific raster data", so I
>> suppose I can answer my own question that teem should be useful. But
>> I don't see any mention in the mailing list archive (assuming the
>> search is working).
>> Also also, mailing list traffic seems low for 2012 and especially
>> 2013. Is the project falling out of use, or just stable and good at
>> what it does with no new features lately?
>> Martin Muc
>> P.S. are you aware there is another unrelated project called teem?
>> (googled "teem usage examples" in trying to answer my own question)
>> and found http://lis2.epfl.ch/resources/download/doc1.0/teem/index.html
>> See everything from the browser to the database with AppDynamics
>> Get end-to-end visibility with application monitoring from AppDynamics
>> Isolate bottlenecks and diagnose root cause in seconds.
>> Start your free trial of AppDynamics Pro today!
>> teem-users mailing list