Deep Forensic Facial Reconstruction (DFFR) - 2 - Futurelearn Course
FutureLearn
The Course from future learn as i noted in a previous post can be free to audit for 14 days.
Week 1 of 2 - The story of bones
This is an intro course and is really meant to give a broad , simplistic look at forensic deduction in week 1. An outline of the "Mr. X" case is given, certain participants from the actual case are used to explain their roles. By the end the user should be able to tell
- Sex of a skeleton
- Sex by skull
- General ethnicity of Skull based on shape of indicator areas and shape of teeth
- identify habitual indicators (ie. did the person chew to one side, etc.)
- give a WAG of age using teeth (transparency of roots) and hips
By the end of week one the user learns that Mr. X was a male of indo european decent , middle to late aged (45-65) that chewed khat to the point he was missing teeth . A DNA review confirmed that the person was not from the UK. Mr. X also showed indications on the bones of his right arm that he likely flexed that arm constatnly. This might indicate a profession where one would have to flex the right arm often. The condition of the remains (he had become an adipocerous mummy) indicated the body had been stored in a cool moist place for years despite being found "recently" in a gym bag.
Week 2 Intro
When would you need a facial reconstruction. Well as the course states, this is a powerful identification tool when used with a publicity campaign. I'd like to note that crowd sourcing is powerful (one example of crowd sourced cyber slueths - billy jenson's "Chase darkness with me") tool. In the context of a GAN generated face , facial recognition becomes a digital version of crowd sourcing.
Back on topic, the Facelab was contacted to help with the Mr. X case. I consider the Facelab group to be the cream of the crop in digital facial reconstruction. Their results are amazing. Their process uses computer graphics software, 3d software, and uses haptic feed back to let a user craft as they would with clay, but in 3d. From what i can tell there is little that is automated or generated by the software.
The Facelab rep in the course points out they are usually brought in when dealing with a cold case or an open case that has hit a dead end.
"It's generally 2 days work to create a finished face"
Go back to the 1000 UHR that stay unidentified per year. 2000 days of labor per year to create a facial reconstruction for each person. That is 2000 days of labor in a lab where it is not clear how many people are working at one time.