Polarik again on a Technical Roll

Here's a cool technical exchange taking place over at the Intrade forum, in the CertifiGate thread that I set up a while back. Some troll logged in as jbeyer posted trashtalk about Polarik and so Polarik has logged on and proceeded to hand him his hat.

ko, I've forgotten to mention to you that Polarik's report is a bunch of hogwash. He appears to be a computer nerd who has written a paper that he knows most people won't understand. His paper provides more "proof" for everyone wearing tin-foil hats. I work in computer graphics, I understand what he is saying, and I know it is a bunch of speculation.

There are rigorous, accepted techniques for detecting digital forgeries, but Polarik has conveniently not included the results of those tests in his paper. If he really believed it was a forgery, he could run some of the tests described in the following papers to prove it was a forgery: ftp://publications.ai.mit.edu/ai-publications/pdf/AIM-1657.pdf http://www.cs.dartmouth.edu/farid/publications/ih04.pdf

Out of curiosity, do you think factcheck.org has been duped or do you believe that they are a part of this vast conspiracy theory?

Joined: 23/09/2008 15:26:34 Messages: 113 Online Hi Ko, just to elaborate on why I feel that it's pointless for lay people (probably most of us) to throw around apparently "expert" opinion as evidence:

(a) Polarik's original report: http://www.freerepublic.com/focus/f-bloggers/2136816/posts

(b) Krawetz debunking Polarik's creditials and methods in (a): http://hackerfactor.com/blog/index.php?/categories/13-Politics

(c) Polarik debunking Krawetz's creditials and methods in (b): http://bogusbirthcertificate.blogspot.com/2008/12/bad-stridence-proof-positive-that.html

This is only one example. See how pointless it is to throw out expert analyses as evidence? This is such a contentious issue that for every piece of evidence you or anyone throws out, someone is likely to debunk it. In the end it boils down to opinion, opinion, opinion.

And in this case the only opinion that will really matter is not yours or mine but that of the courts, assuming they will take the case up. But that doesn't seem likely, does it? Which is why I said that if this contract is put up for trading, it will trade ridiculously close to "not forged".

Out of curiosity, do you think factcheck.org has been duped or do you believe that they are a part of this vast conspiracy theory?

What a bunch of garbage this is, coming from yet another FRAUD who has NEVER READ my 160-page final report. If you had read it, with a modicum of comprehension, then you would know why both the COLB scan image and the photographs were never made from Obama's real Certification of Live Birth.

You would also know that a whole bunch of readers DO understand what i wrote because i took a visual, pragmatic approach with 140 images for illustration purposes, and explained what I did in common sense terms, wherever possible.

If you knew anything about the esoteric procedures you conveniently did not discuss, then you would know why they could NOT be used, nor would I bother to use them, because then it really would take a high-level expert to understand them.

First of all, it doesn't take a rocket scientist, or a puffed-up computer expert, to know that when the Seal on Obama's 2007 COLB does not match the real Seal used for 2007, but does match a different year, you're looking at a stone-cold forgery.

Secondly, I wish I had a dollar for how many self-professed computer "experts" have said that the pixel patterns I found are "scanner artifacts," or "JPG artifacts."

Instead of debunking my research, they reveal their true ignorance of scanner graphics.

The visual, pragmatic, trial-and-error approach I used, and explained, rules out any naturally-occurring phenomena which could have produced the anomalies I found.

If you're so smart, why didn't you explain these techniques to us instead of just throwing up a link to this research article (that was done primarily for publication purposes)?

Why not show us who else has used them in a practical application of identifying a fraudulent scanner image?

Better yet, show us how they can be used to analyze a scanner image when you don't have the original scan for comparison purposes, as required by these techniques?

Well, you can't. Frauds love to use big words and list esoteric techniques hoping that nobody else knows about them.

It is not only being disingenuous, but you and others like you are always trying to set me up like this. Conversely, it is absolutely ludicrous to do so because there is no valid argument that anyone can postulate as to why, after six months, all we've ever seen is only ONE SUSPICIOUS-LOOKING SCAN OF THE FRONT SIDE of a document that allegedly exists (but does not).

Nobody who believes that this COLB scan is genuine has ever done anything to prove that it is genuine. All they have done is to say, "It looks genuine to me."

What a crock! Seriously! These fools will disregard anything I tell them, regardless of the method i use. They will still falsely accuse me of messing with the evidence. They will also counter that I have not seen the original document -- yeah, and nobody will because it does not exist.

The reason why anyone would use these techniques is because they cannot visually detect digital alterations in a photo and they have the original photo to which comparisons can be made.

These techniques do not replace the human eye, nor do they take the place of pragmatic visual analysis. They cannot say HOW an image was altered -- the most that they can say is that an image MAY have been resaved, recompressed, or spliced together from two or more other images.

Secondly, the formulas and assumptions they use for detecting resampling will only work with images saved with linear resampling algorithms, such as nearest neighbor. If the original image was saved using a nonlinear algorithm, then these formulas will fail to identify true forgeries:

"It is not possible, however, to uniquely determine the specific amount of re-sampling as there are re-sampling parameters that yield similar periodic patterns. There is also a range of re-sampling rates that will not introduce periodic correlations."

Their routines for detecting double JPEG compression only work when the save-to-save compression rates are known AND are different AND the second rate is not a multiple of the first, AND when there are no artifacts in the original image that would confound the results.

Since my detractors insist that the original scan has naturally-occurring "scanner artifacts," and "JPG artifacts," then, if true, that would rule out using these techniques to convince them.

Same thing for luminescence changes: you need to have the original, but you can tell by looking at the histograms which parts of the forged COLB image very widely in terms of hue and saturation; the border, for example -- which I have shown could not have been made at the same time that the rest of the scan was made.

The signal-to-noise procedure is also limited to having an original image or detecting which part of an altered image was not altered. If noise was added to the whole image, which is what was done to the forged COLB image, then there is no way to identify any deviations from expected parameters without having the original image to provide them:

"Note that this estimator assumes a known kurtosis for the original signal and the noise, kx and kw. In general these quantities may not be known. In the results presented below, we assume that they are known. In the future, the kurtosis of the original signal can be estimated from a region of an image that is believed to be authentic, and the kurtosis of the noise can be estimated by, for example, assuming Gaussian noise (kw = 3), or modeling the noise statistics of JPEG compression."

The intent of my research was to identify how the anomalies in the COLB image were produced, by trying to replicate them using natural and man-made processes. The inherent advantage in my approach is that it demonstrated how they were produced while simultaneously eliminating the processes that critics claimed were responsible for their appearance.

The bottom line is that what was posted online is a bogus Certification of Live Birth image, and so far, no one has been able to prove otherwise.

Joined: 11/10/2008 03:32:14 Messages: 40 Online Polarik, PhD, I don't know if you'll read through my whole post, so I'll post the question that I'm most interested in getting a response to first. I don't mean to demean you, your education or your knowledge, but am earnestly interested in a response to this question. If this birth certificate is so patently a forgery, why were you the best expert witness that Alan Keyes could find to testify? The credibility of an expert witness is almost solely based on their reputation. You were forced to write your opinion anonymously, which seriously undermines your credibility. Alan Keyes and his lawyers were well aware of this. This suggests that they couldn't find AYNONE with a reasonable reputation to support the forgery theory.

What a bunch of garbage this is, coming from yet another FRAUD who has NEVER READ my 160-page final report.

Polarik, PhD, I'm sorry if I touched a nerve or two. Polarik, PhD, why do you assume that I haven't read your report? It was painful, but I read every word of it. I wouldn't call something hogwash if I hadn't read it and understood it.

I'm not going to get into a credential-throwing fight here. Your PhD in instructional media probably has my PhD in AI beat, but I am familiar with the procedures that I linked to.

Ever heard of a program I use called, "JMicrovision?" Probably not - otherwise you would not have simply tossed out a link here. I initially used it for analysis purposes, but because nobody except a computer expert would understand it, I chose a different approach -- one that others could understand....If you're so smart, why didn't you explain these techniques to us instead of just throwing up a link to this research article (that was done primarily for publication purposes)?

I didn't explain Farid's techniques in this forum for the same reason that you didn't write an analysis that only computer experts could understand. How many computer experts are in this forum and could have understood what I wrote?

Keep the tin-foil hat with you at all times, lest myself or others like me come after you.

Nobody who believes that this COLB scan is genuine has ever done anything to prove that it is genuine.

I don't really feel that I need to prove it is genuine when the Director of Health of Hawaii has issued a press release stating it is genuine. It is generally the job of conspiracy theorists to prove their theory, and not the other way around.

I don't believe that you messed with evidence, I just don't believe you have shown much valid evidence.

They cannot say HOW an image was altered -- the most that they can say is that an image MAY have been resaved, recompressed, or spliced together from two or more other images.

Then run the bloody test and show that it has been altered. If you ran these tests and showed, using peer-reviewed techniques, that the scan had been altered, that would be something you can hang your hat on. But to say that you won't run the tests because it will only show you that it has been altered, and not how it has been altered, is just idiotic.

That is like saying that you won't check to see if a patient is dead unless you can also figure out why the patient died.

ko, I've forgotten to mention to you that Polarik's report is a bunch of hogwash.

What a bunch of garbage this is, coming from yet another FRAUD who has NEVER READ my 160-page final report.

Well, it looks like jb stepped into a big, steaming pile here This thread just got a heckuva lot more interesting.

While I'm sure that ko(razy) won't believe me, I work in the computer graphics industry, and can understand both Krawetz and Polarik's reports.

I guess we'll all see about that, now that Polarik just ripped you a new one right here on this thread.

Too bad you burned up all that good will by going straight for the troll act, I might have actually been motivated to answer your questions to me.

Hmmm....
I wonder just who was Obama's mother?
Could Ayer's and Crew slipped someone into place....I know it is farfetched...but why Obama won't reveal important info is just Crazy....

I have a baby book with photos of my childhood; I have a birth certificate; I have school mates and former teachers and professors who recall me; I have all this and much more which Hussein does not have.

Various lists have been assembled showing the dark mass missing from the Universe of Hussein, greater by far than the mass identified.

A black hole is a fountain of light and information compared to the Hussein record.

Davis was a Communist, a member of CPUSA, a pervert, pedophile and close friend of Stanley--does a Communist have "friends" outside the cell.

Who paid the 75K for Harvard. What passport got Hussein into Pakistan in 1981.

One could ask a hundred such questions and have only the hollow assurances of Hussein's sheepdippers for answer.

Ayers was in Caracas in 2006 praising Chavez, shouting viva la revolucion--having been part of the crew calling for the elimination of the 25M who could not be reeducated per the confidential informant.

Klonsky in Beijing.

The Islamo-Commie who was "born" in two Hawaiian hospitals.

The Columbia graduate no one remembers.

The resident of a mansion bought with Saddam's Oil for Food scam.

John Brennan's The Analysis Corp. illegally accessed Hussein's passport files a year ago when Brennan donated 2K to Obama, but Brennan wants no part of DCI.

What rough beast slouches toward Bethlehem--one who bars its name from speech--

Who

Sane

Oh, gosh no. It was not meant for you. I'm sorry you got the wrong impression.

Thanks for clarifying that, Polarik.

First of all, I have never identified myself as a "forensics expert" on my Affidavits as there is no need to do so. There is no formal field known as "document image forensics," and therefore, no such thing as a "certified expert in document image forensics."

Are your affidavits being used in any of the pending litigation?

There's a well established field called Questionable Documents (or is it "questioned?"), for which there are a couple widely recognized certifications, but I believe its focus is still largely on hard copy, handwriting analysis, etc. The type of analysis at hand would probably be best described as a hybrid of QD and digital image forensics, which also has a lot of overlap with the forensic video field.

What I was saying by my comment is that "Forensics" is not synonymous with "CSI"...

Amen.

I know all that. It is just the same old line the trolls trot out. Actually I think he was referring to the statement made by the clerk about the COLB, not the official statement made by the Director about the BC on file

Nope, the link on the forum went to the statement of DR. Chiyome Fukino, the Director of Health for the State of Hawai‘i which was:

“There have been numerous requests for Sen. Barack Hussein Obama’s official birth certificate. State law (Hawai‘i Revised Statutes §338-18) prohibits the release of a certified birth certificate to persons who do not have a tangible interest in the vital record.

“Therefore, I as Director of Health for the State of Hawai‘i, along with the Registrar of Vital Statistics who has statutory authority to oversee and maintain these type of vital records, have personally seen and verified that the Hawai‘i State Department of Health has Sen. Obama’s original birth certificate on record in accordance with state policies and procedures.

“No state official, including Governor Linda Lingle, has ever instructed that this vital record be handled in a manner different from any other vital record in the possession of the State of Hawai‘i.”

Notice she does not say she or registrar has seen the original certificate, but rather that she has "seen and verified" that one exists in the state's records.

I didn't see the link on the forum. Anyway, it still shows the guy doesn't know the basic facts. What the Director was commenting on was his birth certificate, not the online COLB which was the topic of discussion. He was saying she verified the COLB as genuine, which she has never made a comment on. You are correct in the statement she did make :

Fukino said she has “personally seen and verified that the Hawaii State Department of Health has Sen. Obama’s original birth certificate on record in accordance with state policies and procedures.

She did not say original Hawaii birth certificate. Under Hawaiian law, he could have applied for a Hawaii birth certificate provided he could show his mother was a resident of Hawaii one year prior to his birth. Part of that process is submitting your original birth certificate. So therefore, in accordance with state policies and procedures , they would have his original birth certificate on record. It does not mean it is a Hawaii birth certificate.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.