Message Boards Message Boards

Diversity in Anatomical model?

Posted 4 years ago

I was looking at Mathematica anatomical models and was able to generate images like this:

AnatomyPlot3D[{Entity["AnatomicalStructure", "SkinOfNeck"], 
  Entity["AnatomicalStructure", "Ear"], 
  Entity["AnatomicalStructure", "SkinOfHead"], 
  Entity["AnatomicalStructure", "HairOfHead"]}]

enter image description here

What about options for female, non-blonde and non-white head models? I already posted this on Mathematica Stackexchange:

https://mathematica.stackexchange.com/questions/224479/diversity-in-anatomical-model

and the topic was promptly "closed".

POSTED BY: Mike R
9 Replies

Perhaps the AnatomyData reference page should have a short discussion of the issue. It need not be long, but an acknowledgement of the problem may guide people who want to use this functionality of its limits.

Posted 4 years ago

Daniel,

Thank you for looking at this issue. This is a first step, probably one should change the default skin tone and hair to something neutral. But there is still the issue of gender.

Anatomical models can have two applications - precise anatomical data for medical applications and approximate models for general visualization. It appears Mathematica currently fails in both areas in regards to diversity. I checked several organs for medical use, but they don't seem to contain any data, for example: AnatomyData[Entity["AnatomicalStructure", "Uterus"], {"Graphics3D"}] returns {Missing["NotAvailable"]}. For general visualization, I checked ExampleData["Geometry3D"], it only contains one human image - Beethoven! ExampleData[{"Geometry3D", "Beethoven"}]

As a curator of anatomical data Wolfram Research has the responsibility of making sure they are representative of humans. Since these data are hosted online, they can be updated without waiting for a new version. As for general 3D graphical data, there is simply no excuse as thousands of such 3D models exist, many in the public domain.

POSTED BY: Mike R
Posted 4 years ago

Daniel:

Thank you for not only joining the conversation, but for also joining in a constructive way.

It is awesome for Wolfram to acknowledge the need, even if the options are limited at this time.

POSTED BY: Mike Besso

We have been discussing this in-house this weekend. Here is some code that emerged for purposes of changing colors.

Manipulate[
 AnatomyPlot3D[{AnatomyStyling[<|
     Entity["AnatomicalStructure", "SkinOfNeck"] -> Directive[sc],
     Entity["AnatomicalStructure", "SkinOfHead"] -> Directive[sc],
     Entity["AnatomicalStructure", "Ear"] -> Directive[sc],
     Entity["AnatomicalStructure", "HairOfHead"] ->
      Directive[Specularity[GrayLevel[.1], 1], hc]|>],
   Entity["AnatomicalStructure", "SkinOfNeck"],
   Entity["AnatomicalStructure", "Ear"],
   Entity["AnatomicalStructure", "SkinOfHead"],
   Entity["AnatomicalStructure", "HairOfHead"]}],
 {{sc, RGBColor[{0.737, 0.592, 0.384}], "skin color"},
  Dynamic[skincolors, SynchronousUpdating -> False],
  ControlType -> SetterBar},
 {{hc, Brown, "hair color"}, {Black, Gray, White, Brown, Yellow,
   Orange}, ControlType -> SetterBar},
 SynchronousUpdating -> False, SynchronousInitialization -> False,
 Initialization :> (
   getSkinColors[skintype_Integer /; 1 <= skintype <= 6] :=
    Cases[WolframAlpha[
      "skin type " <>
       IntegerString[skintype], {{"BasicProperties:SkinTypes", 1},
       "Content"},
      PodStates -> {"BasicProperties:SkinTypes__More",
        "BasicProperties:SkinTypes__Less"}], _RGBColor, \[Infinity]];
   skincolors = getSkinColors[5]
   )] 

An alternative:

getSkinColors[skintype_Integer /; 1 <= skintype <= 6] :=
 Cases[WolframAlpha[
   "skin type " <>
    IntegerString[skintype], {{"BasicProperties:SkinTypes", 1},
    "Content"},
   PodStates -> {"BasicProperties:SkinTypes__More",
     "BasicProperties:SkinTypes__Less"}], _RGBColor, \[Infinity]]

skincolors = getSkinColors[5]

AnatomyPlot3D[{AnatomyStyling[<|
    Entity["AnatomicalStructure", "SkinOfNeck"] ->
     Directive[skincolors[[7]]],
    Entity["AnatomicalStructure", "SkinOfHead"] ->
     Directive[skincolors[[7]]],
    Entity["AnatomicalStructure", "Ear"] ->
     Directive[skincolors[[7]]],
    Entity["AnatomicalStructure", "HairOfHead"] ->
     Directive[Specularity[GrayLevel[.1], 1], GrayLevel[.1]]|>],
  Entity["AnatomicalStructure", "SkinOfNeck"],
  Entity["AnatomicalStructure", "Ear"],
  Entity["AnatomicalStructure", "SkinOfHead"],
  Entity["AnatomicalStructure", "HairOfHead"]}]

(I am no expert in this area and am posting what has been proposed. We recognize this is far from perfect.)

As for the lack of diversity in the available model, I will give my understanding of the present status. Currently our one access is to a free source, from a project that is not heavily staffed. They are aware there is a need for far greater diversity in terms of age, gender, geography and ethnicity. But we do not have a time frame for when more such data might become available to us. We apologize for any inconvenience that this might cause for anatomical modeling or other applications

POSTED BY: Daniel Lichtblau
Posted 4 years ago

I would not expect a data curator like Wolfram to manipulate content.

That said, I think it would be a great feature to allow end users to adjust the colors.

POSTED BY: Mike Besso

I'm not entirely buying this. 50%+ of the population are women. Light-skinned people make up a minority of the world's population, and blondes make up only a small fraction.

Unless prevented from doing so by license restrictions, WRI could edit the figures to have (at a minimum) brown/black hair and a more neutral skin color. It would not address the larger issue, but at the very least would not be quite so offensive.

It would also be useful to provide a way to change colors, but this may be available already. If The Sims can do it, why not Wolfram?

This is and issue in general with medical data, and is only slowly being addressed. Part of the problem is that to account for gender, race, and age differences literally makes the study design and analysis at least 25 times harder. With limited budgets to conduct studies, investigators focus on the biggest group(s).

POSTED BY: Robert Nachbar
Posted 4 years ago

I think it is great that we are having this discussion.

George makes a great point. Wolfram, and many other curators, rely on outside sources.

Perhaps we should also be making this point with the outside curators. But, even so, it will take time to see the needed change. While patience is a virtue, we must continue to demand progress.

Thank you for making me aware of this. I hope some of you join me in continuing to raise the need for diversity and inclusion with content makers.

POSTED BY: Mike Besso

Also, the skeleton, for the hip especially, is male only. I know that Wolfram is dependent on outside sources for the data they curate, but still....

Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract