Monday, September 12, 2005

9/12/05

5:11 a.m. Back when I was a student I went down to Scotsdale, AZ to learn how to do forensic facial reconstruction from Betty Pat Gatliff. She was a sculptress that worked for years with Dr. Clyde Snow, a Titan in the forensic anthropology world and the man that revolutionized facial reconstruction techniques using tissue depth measurements gathered from the faces of hundreds of cadavers.

Forensic facial reconstruction is pretty much a last-ditch effort that death investigators use to get a skeleton identified. This is because it is equal proportion art and science.

You cannot, for example, tell the shape of lips or a nose from a skull...or the color and style of hair...or eye color and shape. You can't tell whether the person was fat or skinny. You can't tell if they had scars or moles or traumas the left them with distiguishing marks.

What you CAN tell, (with the help of a forensic anthropologist) is the approximate age, race, and gender of the decedent. You can look at the shape of the nasal bones and get a general idea of whether the nose flared or was turned up at the end. And because of tissue depth markers specific to race and gender, you can get a pretty accurate feel for the shape of the face.

So, basically, if you get a bunch of reconstruction artists working on a caste of the same skull, you'll end up with a room full of sculptures that look like cousins. (Interestingly, without a model to work from, people tended to give their reconstructions noses and lips similar to their own.)

At the time, I was only taking this class as research for an idea I had to gather more accurate tissue depth readings. Even back then, actual clay reconstruction was falling rapidly to the wayside in favor of computer-based reconstruction and age progression models.

Anyway, my idea was based on the fact that gathering tissue depths from cadavers creates a whole host of inaccuracies. The most obvious being the rate at which skin dehydrates after death, scewing your results.

And having asked around to friends, family, strangers on the street, I determined that most people wouldn't volunteer to have needles inserted into several locations on their skulls for the sake of my research. My hope, provided I could wade through a ton of beaurocratic red tape, was to use the thousands of cross-secting head CT scans stored in the university hospital's radiology department to get more accurate measurements.

Anyway, my nerd is showing. How embarrassing.

I never ended up undertaking the project for various reasons, but I'm sure it's been done ten times over by now. There is no such thing as an original thought, after all.

7 comments:

PollyME said...

Olga,

Could you please call me, NOW!

Mike Weasel said...

Sure there is. For example, what if it was possible to genetically engineer a giant flying rhinocerous. Then, what if we could travel to Saturn by riding inside its stomach?

Heh? I bet that's an original thought. I just came up with it now.

OK, so it's not a very useful or practical though, but it is original. If I can do it, so can you.

PollyME said...

Weasel,
I came up with the concept of the Rhino Transport-Ocerous YEARS ago back when I was going through my glue-sniffing phase.

Nice try, though.

Stephen Blackmoore said...

Don't bet that it's been done until you see it in the field. Lots of people have the same thought that nobody moves on.

For example, from this article, the quote, "Currently, data collection from CT scans is a laborious and time-consuming process of amassing a series of point-to-point measurements. The automated capture of 3D data from CT scans is an imminent proposition, but will not be cost effective for facial reconstruction in the near future."

I'd be willing to bet that the reason for that is that nobody has bothered to think of a workflow to make it less time consuming.

You've already got your data model. The landmark points have been well established for some time now. You just need the data itself, depth ranges based on gender, and race. it won't do the job for you, but it would give whatever software or sculptor using it better information.

Categorizing the data and creating a series of sets and models that can be easily used across applications could be extraordinarily beneficial. Say, in an XML format that could be read by anyone who wanted to create a software engine to use it.

Granted, you would want to set up a databse to do searches in, since XML takes forever to slog through, but if you came up with a forensics XML schema (I'm sure there's a genetics one - but facial reconstruction?) imagine how much more efficient identification could be.

Okay. Sorry. Done geeking now.

Anonymous said...

In 1971, I wanted to create an artificial muscle that would not be rejected for polio and accident victims. Probably not an original thought, but as an eight year-old it was way beyond my comprehension how to accomplish that.

Mad Scientist said...

Polly - National Institute of Health has free access to CT and MRI of the entire male and female body online. Don't know if that would help at all but I believe the databases exist I bet you just need the software to analyze it like Stephen said.

My original idea which I am sure someone will now take as their own:
I want to isolate the gene from starfish responsible for regeneration and clone it into Human DNA to allow for people to regrow limbs, organs, etc.

Of course that is a lot easier on paper than in practice...

Anonymous said...

I had an idea - nope, it's gone.

oh!

na, gone too.

...dang, and I know it was original because I'd never thought of it before!