[X3D-Public] progress on HAnim skeletons; anatomy/anthropometry strategy discussion

[Don Brutzman]

"Big list" participants:  H-Anim working group is reaching new levels 
and reaching out to Medical working group to build *anatomically correct 
human skeleton, modeled and animatable in X3D*.

Further expertise is sought.  If you are interested in these topics, 
please follow along and consider contributing on the H-Anim or medical 
mail lists.

Have fun with X3D!

On 7/4/2013 10:55 AM, Don Brutzman wrote:
> Summary:  X3D h-anim model diagnostic review, followed by
> asymmetry/anatomy/anthropometry discussion.
>
> The H-Anim working group has made such excellent progress that this
> discussion now rises to include multiple issues relevant to the medical
> working group, cc:ed.
>
> Bottom line up front (BLUF): can medical group help advise the h-anim
> group to build an anatomically correct, complete and medically
> acceptable skeleton?
>
> ==============================================================================
>
>
> 1. Referenced humanoid models are working in Instant Reality,
> View3dscene, H3DViewer and OctagaVS.
>
>      http://www.web3d.org/x3d/content/examples/Basic/HumanoidAnimation/
>      http://www.web3d.org/x3d/content/examples/Basic/HumanoidAnimation/HAnimSpecificationLOA3Invisible.x3d
>      http://www.web3d.org/x3d/content/examples/Basic/HumanoidAnimation/HAnimSpecificationLOA3Ilustrated.x3d
>      http://www.web3d.org/x3d/content/examples/Basic/HumanoidAnimation/HAnimSpecificationLOA3Motion.x3d
>      http://www.web3d.org/x3d/content/examples/Basic/HumanoidAnimation/HAnimSpecificationLOA3Motion.html
>
> ==============================================================================
>
>
> 2.  The X3D Schematron report from yesterday's H-Anim working group
> teleconference included the following warning:
>
>> <HAnimJoint DEF='hanim_HumanoidRoot'/> has incoming <ROUTE
>> toNode='hanim_HumanoidRoot' toField='translation'/> to modify the
>> translation field, which ordinarily is not modified (instead the
>> center field controls HAnimJoint position)
>> [/X3D/Scene/HAnimHumanoid/HAnimJoint[1], warning]
>
> As you pointed out at the end of the call, Joe, the HAnimJoint named
> HumanoidRoot is a special case that actually does get animated via the
> translation field to move the humanoid.  Agreed, that is a key point.
> Also as discussed, none of the other HAnimJoint nodes are intended to be
> animated in this way.
>
> Therefore I have relaxed the general HAnimJoint warning rule to not
> report on this special case.  It will be deployed in next week's build.
>   So that false-positive error is all set.
>
> Anyone can perform the validation tests on these scenes, using the
> online X3D Validator.  Attached screenshots illustrate how (click the
> highlighted checkmark button on the Basic Examples archive web page).
>
> This leaves us with only one class of reported warnings on our improved
> skeleton model, as follows.
>
> ==============================================================================
>
>
> 3.  The remaining 31 scene warnings are all due to joint asymmetries: l_
> left joint center values not being exactly opposite the r_ joint center
> values.
>
> I spot-checked the first three warnings, and each pair of center values
> being flagged are an exact match for the values given in the H-Anim
> specification:
>
>      Annex A (informative) Nominal body dimensions and levels of
> articulation
>      http://www.web3d.org/files/specifications/19774/V1.0/HAnim/BodyDimensionsAndLOAs.html
>
> There is nothing wrong with a body being asymmetric (many of us are,
> cough cough) but asymmetric values don't seem like a good default model
> for us to use, especially when trying to build a nominal generic
> humanoid that is free of errors and anomalies.
>
> Asymmetry should be an author choice - and then providing warnings helps
> make that choice actionable.  In that case, the default values in the
> H-Anim specification need updating.
>
> So I think this brings us back to anthropometry and choosing default
> values that match on opposite sides.  What resource shall we use?
>
> ==============================================================================
>
>
> 4.  William, thanks also for the insightful discussion last night on
> different possible models that might be available.
>
> Wondering, who was working on the anthropometry portion of the h-anim
> working group wiki?  I'm not able to follow what is listed on the wiki
> very well:
>
>      http://www.web3d.org/wiki/index.php/H-Anim
>
>> Published References for review and citation
>>
>> Professional Societies: Professional Certifications, Licenses in
>> Anthropometry International Standards: ISO - Basic human body
>> measurements for technological design -- Part 1: Body measurement
>> definitions and landmarks IEEE -
>>
>> North American Standards:
>>
>> ANSI
>>
>> South, Central, Latin American Standards:
>>
>> European, Middle Eastern and African Standards:
>>
>> Asian Standards:
>
> I added a section title so that we might have a direct link to build out
> this needed collection of Anthropometry information:
>
>      http://www.web3d.org/wiki/index.php/H-Anim#Anthropometry
>
> It was good to discuss this topic for a while last night.  It wouldn't
> surprise me much if you do find different references for different
> places, as listed above.  However, I'm concerned that such a worthwhile
> effort might unfortunately take us in the wrong direction.
>
> Over the years, in the graphics community, several
> confounding/contradictory trends have emerged:
> a. any number of schemes have been proposed for efficient facial
> animation over low-bandwidth network connections.  typically these are
> patented and closely held as part of a company's business portfolio.
> b. CGI animators and motion-tracking professionals build and animate
> models that are wildly divergent in order to meet particular animation
> goals, consistency is not seem as a high priority.
>
> This is all understandable and OK because different design goals and
> technological assets are being applied.  And no doubt these kinds of
> things can continue to be supported/demonstrated by X3D examples.  But
> wild generality doesn't have to be our top priority as a working group.
>
> I think that if we want to have a truly flexible and adaptable H-Anim
> specification, we need to get the "base case" of an anatomically correct
> humanoid working as well.  This can have further major benefits as an
> asset for medical use.
>
> We therefore need to know what is the primary reference for human
> skeletal anatomy, accepted by the medical medical community worldwide,
> and unencumbered by legal impediments.  If we get the authoritative
> "base case" correct, then mappings for variant approaches can be handled
> individually.
>
> Wondering, what are the right references to look at - Dick and Mike, can
> you help?
> - ISO standards?
> - Medical community accepted references?
> - Publicly available references?
>
> Perhaps the answer is quite simple: just use Wikipedia.  That is free,
> widely accessible, steadily scrutinized for correctness, easily
> referencable, not restricted to medical experts, translated into many
> many human languages, and not confounded by potentially competing or
> contradictory nomenclatures.
>
> Example articles:
> - http://en.wikipedia.org/wiki/Human_skeleton
> - http://en.wikipedia.org/wiki/Axial_skeleton
> - http://en.wikipedia.org/wiki/Appendicular_skeleton
> - http://en.wikipedia.org/wiki/Human_vertebral_column
> - etc.
>
> These look to give us names and relationships of bones in the skeleton,
> but do not appear to give typical linear/angular dimensions or
> measurements.  The article on anthropometry says almost nothing about
> skeletons.  Kinanthropometry is another possibility that points towards
> sports medicine and clothing, among other topics.
> - http://en.wikipedia.org/wiki/Anthropometry
> - http://en.wikipedia.org/wiki/Kinanthropometry
>
> A good index of Wikipedia assets follows.  Still not finding dimensions
> anywhere... can anyone else?
> - http://en.wikipedia.org/wiki/Category:Human_anatomy
>
> Several things to think about here, but with a pretty clear and simple
> goal:  *anatomically correct human skeleton, modeled and animatable in
> X3D*.
>
> Suggested path forward for h-anim group, guided by medical group:
> - continued discussion and literature review, especially from experts
> - decide on preferred anatomical references
> - define additional H-Anim Levels of Articulation (LOAs) to match a
> fully defined skeleton
> - build test and use X3D examples
>
> All feedback welcome.  For those celebrating in U.S., Happy Independence
> Day.

all the best, Don
-- 
Don Brutzman  Naval Postgraduate School, Code USW/Br       brutzman at nps.edu
Watkins 270,  MOVES Institute, Monterey CA 93943-5000 USA   +1.831.656.2149
X3D graphics, virtual worlds, navy robotics http://faculty.nps.edu/brutzman