AA/FS Photo Comparison - Similarities: What is Wrong With AA Being FS?

[AGRBear]

I mentioned pierced ears on another thread and this was Elisabeth's response which gives us more information on the photographs which shows FS wearing earrings.

Bear...

The photo you are now citing as evidence that FS might have had pierced ears was published in a Berlin newspaper and was so obviously and badly retouched that not only does FS look like a cartoon character in it, but the "artist" (if you can call him that) even added earrings where there were none in the original!

So we simply don't know if FS had pierced ears or not. The whole issue is a red herring. And since we have already discussed this at some length on another thread, I have to wonder why are you bringing it up again here when you know for a fact that this retouched photo was in error?
 

By the way, I was not saying that FS had pierced ears in the post on the other thread Elisabeth was responding, I said that particular  photo, which in an earlier post I said was touched by an artists or someone,   showed she had pierced ears which "may be in error" ....  From the photo that is close to the original, we can't tell if she is wearing earrings or not.

AGRBear

[Helen_Azar]

This is one of Helen's photographs.  She wrote under it
"Un-retouched version of the photo^^"

Bear, please do not misquote me. I did not write that this was the unretouched photo of FS. In fact, I described this particular version as a very bad artist rendition of the FS photo.  Once again, I would appreciate it if you didn't put in my mouth as you often tend to do, for some reason.

This was my original posting:

I just found this FS photo which is supposedly the original version - before it was manipulated (next to it are the two "re-touched" versions):


Un-retouched version of the photo^^

The first re-touched version doesn't look that different to me, only the features are more clear. It doesn't really look as much as it was re-touched as an attempt to make the photo more clear, this was probably the purpose for it. The second picture just looks like a bad artist's rendition, and is neither here nor there...

There are certain photos of AA, not all, that bear a striking resemblance to this FS photo (unre-touched). What do you think?



Fraziska Schankowzska (unre-touched photo)                      Anna Anderson                                                       Anna Anderson

To be honest, even based on these three photos alone , I would say that there is a really good chance that this is one and the same person - I mean the resemblance is really uncanny here!

[jaa]

I had no idea the image was so blurry.

Part of that blurriness is caused by generational loss; a resized digital image of a cropped and enlarged photograph printed in a book. But the original would still be grainy and blurry, given the types of film, lenses, and photographic paper used in everyday photography in 1916. AGRBear posted the mostly uncropped image:


The image used in the documentary appears to be clearer, but I can't tell if it is or if it's just picking up details from the underlying AA image. I wish they had completed the fade-in.

If it is clearer, it might have been image-enhanced. I hope they kept the first high-resolution scan. The program is copyright 1995, and there have been significant improvements in image enhancement technology in the last ten years.

Also many, many thanks for taking the time to explain the technical side of retouching photos - the process is a lot clearer to me now. No wonder they call it an art!

Thanks for the thanks (Penny and AGRBear too). I don't know what details would be of interest to a historian, but I thought I'd try to contribute what I know about the technical side.

One thing I don't understand about this dissolve is the rotation of the FS photo. Rotating a bit-mapped image on a computer causes a certain amount of distortion. This is because the software has to interpolate the new position and color of each pixel. I wonder why they didn't rotate the AA photo instead since it contains far more data and would therefore suffer less distortion. I'm not accusing anyone of falsifying anything, BTW. Maybe it's what would look better, since documentaries are a form of entertainment.

[AGRBear]

Helen,

Unfortunately, you didn't read through all of the posts.  I already apologized for misreading your original post which had under the retouched photo the following words:  "Un-retouched version of the photo^^ ""  The two "^^"  were pointing at the retouched photo.  Just as they are, now, above in your post.

This is the "RETOUCHED" photo of FS.


"Un-retouched version of the photo^^"
[words which on my screen shows under the above photo]

Above the three photographs were these words which I obviously didn't read or didn't read it correctly where Helen explained the following three photographs:  >>I just found this FS photo which is supposedly the original version - before it was manipulated (next to it are the two "re-touched" versions): <<

Everyone, again, let me say:  I mistook the words under the photo as meaning it was pointing at that photo.  Sorry, again, Helen.

As far as I know, I have never tried to place words in Helen's mouth.  If I did, let me go and correct it immediately.

AGRBear

NOTE:  1 April 2005 -  One mystery solved,  my screen shows Helen's photos one on top of the other
X
X
X

And Helen's screen shows her photos in a row
X X X

Therein is the difference and the difference of the  position of the "^^" .....

So, a lesson learned.  Remember what you see on your screen may be different then what others see

[Helen_Azar]

Helen,

Unfortunately, you didn't read through all of the posts.  I already apologized for misreading your original post which had under the retouched photo the following words:  "Un-retouched version of the photo^^ ""  The two "^^"  were pointing at the retouched photo.  

Bear, apology accepted. But please look closely at my original post again, and you will see that the arrows ^^ are pointing towards a different photo than the one you posted and quoted me as saying it was "unretouched'.  My arrows are pointing towards the actual unretouched photo and not the one I identified as a bad artist's rendition (it is not even a retouched photo) as you claimed.
Please be careful of how you quote people.

Un-retouched version of the photo^^

[AGRBear]

..[in part]...

What AA photo do I place next to FS which was used in the Nova's tv show for the overlay?


AGRBear

[AGRBear]

FS>

[second photo is absent because it was changed by owner of photo to something else....]



AGRBear

PS  The FS is still a copy of a copy which shows the two white spots [perhaps lint which occurred during the copying of the original]  in her hair...  And AA's is also a copy and I don't know the origin....

[jaa]

Thanks for doing that, AGRBear. Sorry for derailing the thread.

I should also say that it's possible to re-create this 1995 dissolve with most image and video editors. If anyone wants to try it in PhotoShop, place the images on different layers and use the move, rotate, and resize tools to align the images. Then use ImageReady and tween the frames between the images.

Just a bit more about terminology, then I'll go back to lurking.

"Image fusion" refers to creating a composite, or combined, image from two different images. Artists use image fusion to create surrealistic images, eg. a composite image of a man and a tree. In mapping and the military, the images are of the same subject, usually under different lighting conditions, such as daylight and infrared. Image fusion is used extensively in medical imaging, such as creating a composite of a patient's CT scan and MRI.

[sokolova]

I suppose I could be wrong - but as I remember it  (I saw the program), the  perfect match was not with AA and FS but with AA and AN!

Something is wrong here. Either my memory or someone's reporting. I have the program on video, so I will watch it again tonight and report the results.

This is all so fascinating to me!

Sokolova

[Abby]

Hi, Sokolova! I beleive you are thinking of the 'ear' test, which they performed in the same program as the 'fusion of the photos' test. When they compared Anna Anderson's right ear with that of Anastasia's, the ears matched. (The expert gave it a '4' on a scale of 1-5, 5 being a perfect match. He would've given the match a 5, he said, if the angles of the photos were identical, but they weren't).
The fusion photo test compared the photographed face of AA with FS's, and all the points on the computer screen matched up.

[Helen_Azar]

Yes, Abby is right. They did perform two visual tests on the program: one was the ear test between AA and AN and another was the photo fusion between AA and FS. Both gave matching results. I must point out that neither one of the tests is considered scientific like DNA, since both results are open to interpretation, unlike DNA.

So if I had to put my life on the line for either result, I wouldn't  . The only results we can completely trust in this case are the DNA results  .

H

[jaa]

Sorry to be so technical, but what was in the NOVA documentary is not image fusion. Fusion of the AA and FS images would be image manipulation.

This is an example of fusion between a Noh mask and a Hellenistic or Roman bronze, from a software tutorial:


All that was done with the images on the NOVA documentary was a film transition called a dissolve. Fade in, fade out with the images overlapping.

Dissolves are a very old transition, originating with silent film. This is a cinematic technique. As far as I know, dissolves are not usually used for any kind of scientific imaging.

The technology used today would be face recognition. The copyright date of the documentary is 1995, at which time face recognition was in the early stages of development. Today, it is being used and tested in venues like airports to help identify terrorists and other security risks. Computer-based face recognition relies on a mathematical description of points on the face.

[AGRBear]

The comparison of faces on the TV show had to do with two photographs which, I assume, were manipulated to show the angle of the faces to be the same because they were not in the originals if the two photos above in my post are the ones Geoff ____ [forgot his last name] used in Nova.

Geoff ____ used his pointer to show the points which seem to match in the process he was showing the TV viewer.

But were they [angles of the faces]  changed when the angle of the faces were turned for the fusion of the photographs?

Here is Vasily Flatov and Alexei in the book THE ESCAPE OF ALEXEI, Son of Tsar Nicholas II by Petrov, Lysenho and Egorov who like Geoff was using the "points" of a face to show similarities on pps. 130 and 131.





Not many of us are convinced by these photographs that Vasily Flatov was Alexei.  Even though the "points" seem to be the same.

So, I'm wondering, about a number of things.

I think I'd like to see someone do a new fusion with our modern software which is used today.  And, I'd like to know which photograph of AA is being used.

Also, what are the numbers on how many people faces when  using points are similar and if it's even a good method of using accept in the process of elimination.

AGRBear

[jaa]

But were they [angels of the faces]  changed when the angel of the faces were turned for the fusion of the photographs?

The FS image was rotated. Take a photograph, lay it on a table and spin it around. That's rotation. However, on a computer there is a chance of distortion when images are rotated, so steps have to be taken to minimize the distortion.

Sorry to harp on this, but this is not fusion. Fusion is similar to morphing. If you took a photo of AA and FS and morphed a composite image, that would be fusion.

Fusion would be inherently deceptive in this case.

All that was shown in the NOVA documentary was a dissolve, a basic transition used in movies and television. As far as I can tell, the computer only replaced the photos and transparencies that an expert would use manually. And it added a dissolve. Here is a simple dissolve, part of a PhotoShop tutorial:

Also, what are the numbers on how many peoplefaces when  using points are similar and if it's even a good method of using accept in the process of elimination.

I know facial recognition is now being used at airports and sporting events to spot terrorists, and in prisons, banks, and other secure facilities. One proposed future use for face recognition, as costs go down, is a replacement for PIN numbers at ATMs.

We'll know a bit more about these systems in the fall, when the US government publishes its findings on its latest face recognition vendor comparison. The comparison will take place in August or September.

I suspect that a face recognition system will find that AA looks like FS. And that she looks like AN. This case has so many bizarre coincidences it wouldn't surprise me.

A face recognition system uses nodes, or vector-based points. The nodal points are typically expressed as XYZ coordinates, with X being a horizontal distance, Y a vertical distance, and Z a nearer or further distance. These points are not dots or pixels; vector-based graphics are mathematical and usually have an accuracy of at least fourteen significant digits of precision. That's why, for example, a circle drawn in a raster-based paint program will look jagged when zoomed in, but will look round in a CAD program regardless of the zoom factor.

This image, from MIT's site, shows nodal points used for face recognition. Typically, face recognition systems use a dozen or more nodes, but only four are shown here: the pupils of the eyes, the tip of the nose, and the midpoint of the mouth.


As for the images of Alexei and Filatov, there are lots of inaccuracies. Look at the line ruled through Alexei's eyes - tangent to the outer part of the bottom right lid and the inner part of the left. There's a similar problem with the line ruled on the photo of Filatov. Look at the lines for the included angle of the faces; the lines on the right side of the faces are tangent to the cheekbone and jaw, but the other lines are tangent only to the sideburns and have no other point of tangency. Also the line ruled through Alexei's mouth is off; it should be slightly lower. IMHO the images are so full of drafting errors that they're pretty much useless.

[AGRBear]

I just love jaa's posts.  They are so full of information.  He knows his subject and what a great image of a green apple with a red apple  [or is it red with a green ).  Anyway, let me say, thanks.

And, you're right the points for Alexei and V. F.  don't match even though them seem to match when you glance at them.

And, yes,  it does appear that FS's  and AA's faces are the same in the tv series but knowing what I know about how objects can be distorted during the rotation by the older software, that is why I questioned this process in the first place.

Again, thanks for all the information, jaa.

AGRBear