SOMERTON MAN MYSTERY

The Evidence The Facts In Detail In Depth

SOMERTON MAN MYSTERY: BRINGING THINGS TO A HEAD > THE MX FACE COMPARISONS...NO MATCHES RECORDED....UPDATED December 11th

...THE MX FACE
'ARTIFICIAL INTELLIGENCE'
FACIAL COMPARISONS...

Comparison 1. Charles Webb Vs The Somerton Man Post-Autopsy Photograph:


NO MATCH


Comparison 2. Charles Webb Vs the Pre-Bust Photograph


NO MATCH

Comparison 3. Charles Webb Vs The Somerton Man Bust Photograph:


NO MATCH

Recently I posted a comment on this blog regarding claims made by someone known as David Morgan. David is a fairly recent arrival on the Somerton Man mystery forums and has set out to make a name for himself as something of an expert especially when it comes to the use of Artificial Intelligence Facial Recognition and Comparison tools. He made the comment on Pete Bowes blog just a day or so ago.

The claim he made was that he had analysed the various faces, the ones shown above, using an unnamed comparison tool. In my case I used a toolset that has a good reputation, MXFace, an online tool that allows anyone to test out comparisons.

Having a little experience in the field of AI, I decided that rather than just accept the word of a newcomer, I would test out the tools myself. It is a straightforward process, here's the link to MXFACE website demonstration, it's a free tool so there's no expense involved:


The results are shown above. In all three comparisons, no match was recorded. That leaves us with some interesting questions.

1. Did I use the tools incorrectly? I don't think so, as you will find they are very straightforward.

2. Did David Morgan make a mistake when he tested the comparisons of MXFACE tools?

3. Is the person who posted the comment on Pete Bowes's blog, not really David Morgan?

Given that I managed to use the tools correctly, I wondered how David got his results. So I followed the link that he? had posted to view the images and results. Below are the images that David used:

This first image is far from being of the quality you would need to get a decent comparison. It's grainy and fuzzy, In fact, I called this one 'Fuzzy'. It was from the large family gathering and not the small family group.







This next image that you see on the right is a young Carl Webb I think, David hadn't named them.






And so we get to the third image. This was part of an experiment by Professor Abbott some 7 years ago or it could be longer. In the experiment, they took a series of photographs of the plaster bust for a 3D clip. This was grab taken from that clip. Not sure why David elected not to use the original Plaster bust as the comparison, maybe he can answer that.





In the interest of fairness, I have run comparisons of the younger Charles as well as the 'Fuzzy Charles' for you to make your own assessments I have used the photographs found on the link provided by David Morgan

1. Fuzzy image of Charles Vs the Somerton Man Post Autopsy:


NO MATCH

2. Fuzzy Image of Charles Vs the Pre-Bust photograph:


NO MATCH

3. Fuzzy Image of Charles Vs the Original Plaster Bust:


NO MATCH
 4. Young Image of Carl Vs the Post-Autopsy photograph:


NO MATCH
5. Charles Richard Web b Vs Post -Autopsy Photograph:


NO MATCH

FINALLY A PROOF TEST:

1. Charles Webb Vs Charles Webb:


MATCHED
2. Roy Webb Vs Roy Webb In Uniform


MATCH

GALLERY

Here's a gallery of all the photographs used in this comparison exercise, please feel free to use them.


This image sourced from ABC News
This image was sourced from the download link shown above.










This image sourced from NAA file
I had a copy of this image in my files
This image was sourced from the download link provided by David Morgan earlier in this post
This image from SA Police files

This image from the SA Police files













This image was sourced from my own files










I trust that this post will be of use to those interested in the case, it's fairly comprehensive in terms of image comparisons. If you have any questions please drop me a line.

4 Comments

Hi
Welcome to the Tamam Shud Blog, widely regarded as the leading and most trusted fact and evidence-based blog on the Somerton Man case. Please take a moment to review our comment guidelines here:

https://tamamshud.blogspot.com/p/tamam-shud-blog-rules.html

Visit our YouTube channel:
https://www.youtube.com/channel/UCOamLze8PyNDafjjBGGngJQ

  1. What can we assume from this series of images? Walking through the post and from other information published earlier, we know the following:

    1. The Pre Bust photographs are the man of whom Paul Lawson made the bust.

    2. The hair samples that were used for the DNA analysis, came from the bust

    3. The Pre Bust photograph of the man of whom Paul Lawson made the bust does not match the post autopsy photograph of the man found on the beach on December 1st 1948.

    4. Paul Lawson stated that he used the post-autopsy photographs of the face of the man to model the face on the plaster bust.

    5. The man found on the beach is the man on whom Dr Dwyer performed the autopsy

    6. Dr Dwyer also took a Dental Chart of the man on whom he performed the autopsy.

    7. The photograph of Carl does not match the photograph of the Somerton Man, the Left ears are markedly different in shape.

    8. The Young Carls left ear does not match the left ear of the man found on the beach.

    From this information it appears highly likely that Carl Webb is NOT the Somerton Man. He maybe the the man of whom Paul Lawson created the bust but in doing so he used another man's face as the basis of his work.

    ReplyDelete
  2. Comparing the face with the mx face of the young Web and the photos from the morgue give a 100% match. I don't understand why you didn't have a match.

    ReplyDelete
  3. I just tested that again and got no match between the young Carl and the Morgue photograph. I used MX face as I did last time which is recognised as being amongst the best face comparison tools. You didn't mention which tools you used to get the result that you got?

    I don't understand how you got your match.

    ReplyDelete
  4. Here's screencast link to a video I just took of the process on MXFace, it shows the two images prior to matching and then the match return shows up. Both images are available here if you would like to try them.

    https://www.screencast.com/t/YNSYDHESL

    ReplyDelete
Previous Post Next Post
/body>