Ray tracing used for obfuscation?

Hi there,

I am not sure if I am in the right forum or not but I could need some help. Is it possible to perform ray tracing on a grey scale 8 bits GIF, 24 bits JPEG or 24 bits PNG file? If so could someone use ray tracing to make an object look totally out of focus? Is it also possible to reverse the out of focus process when you know the exact used settings: light intensity, light direction and bit scale?

Why am I asking this. I stumbled upon an image, which I believe was tampered by means of ray tracing. I know quite a lot about obfuscation / tampering but this particular image I have never been able to prove any form of tampering. After image processing of this photo using dozens of contrast / edge enhancing and de-blurring techniques I recently noticed a pattern within the obfuscated area. The image consists of 15 x 15 = 225 squares of blocks of 8 x 8 = 64 pixels each. There are 15 horizontal "rows" of 15 squares and 15 vertical "columns" of 15 squares.

The software I used added artificial light to the smudged object from top to bottom at an angle of 3.6 degrees and by going through each row / column and repeating this process 15 consecutive times I was able to literally remove "one layer" of grey pixels of the smudged object and this grey layer was replaced by "one layer"of darker pixels that seem to come forward from the background.

Than I performed the same operation 15 times from left to right but directed the artificial light on the smudged area at an angle of 19.8 degrees.

I repeated this process 33 x 15 = 495 times in both vertical and horizontal direction and ended up with a CLEAR picture without any smudge. I am still working on the analysis and processing of other images with similar obfuscations so I cannot tell you yet what I found.

Are there any "ray tracing" experts who like to share their expertise and answer my few questions?

Is it really a form of ray tracing or did I accidentally used a different technique?

Thanks for your interest,

Greetz,

Sander

in b4 someone says they have no idea what you are saying and this is way over their heads,

but seriously, im curious what these images you are working on are of,

any chance you could post them up here?

always interesting what you have to say, but i too need more to go on.
i learnt about ray tracing in an advanced photo manipulation class i did long ago.

love the work you do with schooling people about nasa's questionable photos, so i would love for you to expand on this topic

reply to post by 1967sander


i am no expert but, what questions do you have?
about ray tracing?
xploder

reply to post by XPLodER


I only need to know whether the technique I used could be ray tracing. Or is it something else? Fact is that a 100% smudged (out of focus) object after processing turned into a 100% clear object. I am however not sure if the process can / may be described / named as ray tracing and I really to be certain as I need it to built my conclusion for my video. I performed the reverse processing on a 2D image and ray tracing is normally only done to create / enhance CGI or 3D games.

Unfortunately I cannot yet tell you what these photos / objects are I am working on.

Greetz,

Sander

reply to post by 1967sander


Are you planning on adding this technique to your repertoire?

But seriously, I don't understand how this image algorithm could be used for obfuscation since all it does is simulate the effects of illumination in an image, creating virtual reflections, refractions and shadows.

reply to post by Chadwickus

simulate the effects of illumination in an image, creating virtual reflections, refractions and shadows.

That's how far I got also by Googling for ray tracing and that's why I ask it in here. This algorithm is unknown to me and i do not have enough knowledge to describe the process. All I see is that the smudge disappears and what is behind the smudge becomes visible. I do not understand this either. It however could explain why no one has ever been able to reveal these smudged objects
by using conventional methods.

Greetz,

Sander

reply to post by 1967sander



Maybe the folks over at CGtalk.com can help out
forums.cgsociety.org...

As far as I know ray tracing is a term used in 3d rendering. Never heard of it in the context of 2d image processing, and not sure how it makes sense in that context.

As stated above ray tracing is primarily a technique used in 3D modelling. It can be used in 2D modelling as well but isn't really for image processing as you describe.

In 3D I suppose you could map a 2D image onto a flat surface and illuminate that but it would offer less than traditional image processing, especially without a height map.

What you have described above has probably lightened the leading edge of a line roughly perpendicular to the set angle of the light and darkened pixels after the next line as well as lightening the pixels inbetween.

So in short, I dont think ray tracing has been used for obfuscation and your technique just happened to work well for the obfuscation present.

reply to post by 1967sander


Why dont you post a link to the image?

Originally posted by 1967sander
Is it possible to perform ray tracing on a grey scale 8 bits GIF, 24 bits JPEG or 24 bits PNG file?

That depends on your definition of "ray tracing". As far as I know it, ray tracing is used to trace all rays from the "camera's" point of view to any point on a 3D scene, calculating the way the light would be affect by any thing it finds on the way to that specific point.

If so could someone use ray tracing to make an object look totally out of focus?

Yes, but only on a 3D program. The image would be mapped on some object on that 3D scene and then a CGI image would be generated from that 3D scene.

Is it also possible to reverse the out of focus process when you know the exact used settings: light intensity, light direction and bit scale?

That's not very likely, as whenever data disappears there is no way of knowing what was really there.

The image consists of 15 x 15 = 225 squares of blocks of 8 x 8 = 64 pixels each. There are 15 horizontal "rows" of 15 squares and 15 vertical "columns" of 15 squares.

The 8 x 8 pixels squares are a result of JPEG compression.

The software I used added artificial light to the smudged object from top to bottom at an angle of 3.6 degrees and by going through each row / column and repeating this process 15 consecutive times I was able to literally remove "one layer" of grey pixels of the smudged object and this grey layer was replaced by "one layer"of darker pixels that seem to come forward from the background.

What software did you use? What you did was just change the colour of the pixels in the way that light algorithm works.

I repeated this process 33 x 15 = 495 times in both vertical and horizontal direction and ended up with a CLEAR picture without any smudge.

You probably ended up with an image that has no relation to whatever was originally blurred, as there is no way of reversing a destructive process.

Are there any "ray tracing" experts who like to share their expertise and answer my few questions?

I'm far from an expert, but I know that we cannot recover destroyed data.

Is it really a form of ray tracing or did I accidentally used a different technique?

I think you used a different technique, and one that does not uncover any tampering.

reply to post by ArMaP


Hmmm interesting theory and explanation but sorry no go!

The image I have in front of me is not a GIF file neither a JPG file but a 27.8 Mb PNG file without any compression. So there are no JPG compression artifacts. Second, the data containg the object is still there but there is a grey layer covering it. The software I am using somehow removes the grey layer and allows the darker pixels behind it to come forward. I performed dozens of tests using different settings. Light intensity and block scale and I always end up with the same object. Depending on the settings showing more or less finer details. The environment that surrounds the objects I processed is not influeced and does not change during the processing in any way. I deliberately save processed images again as a PNG file (24 bbp) with compression "0" so that there are no (JPG) compression artifacts.

Originally posted by 1967sander
The image I have in front of me is not a GIF file neither a JPG file but a 27.8 Mb PNG file without any compression. So there are no JPG compression artifacts

If the image is from a digital camera it might have originally been a JPEG image that was subsequently converted to PNG. Many cameras have a 'Raw' setting to avoid lossy compression.

Originally posted by 1967sander
...deliberately save processed images again as a PNG file (24 bbp) with compression "0" so that there are no (JPG) compression artifacts

PNG uses lossless compression so don't worry about having compression at full.


You should try masking off the area of interest and playing with 'levels' to see if you can achieve the same effect that your 'add illumination' method is.

Originally posted by 1967sander
The image I have in front of me is not a GIF file neither a JPG file but a 27.8 Mb PNG file without any compression. So there are no JPG compression artifacts.

Are you sure that that image was never a JPEG that was saved as a PNG?

Second, the data containg the object is still there but there is a grey layer covering it.

PNG images do not have layers.

The software I am using somehow removes the grey layer and allows the darker pixels behind it to come forward.

Apparently, what you are doing with that software (what software is it, can you tell us?) is darkening some pixels from that area, giving the idea that it is darkening what you call "layer".

The environment that surrounds the objects I processed is not influeced and does not change during the processing in any way.

Some processes can be applied just to some pixels, like those that affect "shadows" or "highlights".

I deliberately save processed images again as a PNG file (24 bbp) with compression "0" so that there are no (JPG) compression artifacts.

Even if you saved it as PNG with compression that wouldn't create JPEG artefacts, as PNG does not use the JPEG algorithm and never creates any artefacts, PNG uses lossless compression.

It would be easier if you could show us the photo, but I understand that it may not be possible.

Originally posted by 1967sander
Hi there,

I am not sure if I am in the right forum or not but I could need some help. Is it possible to perform ray tracing on a grey scale 8 bits GIF, 24 bits JPEG or 24 bits PNG file? If so could someone use ray tracing to make an object look totally out of focus? Is it also possible to reverse the out of focus process when you know the exact used settings: light intensity, light direction and bit scale?

Why am I asking this. I stumbled upon an image, which I believe was tampered by means of ray tracing. I know quite a lot about obfuscation / tampering but this particular image I have never been able to prove any form of tampering. After image processing of this photo using dozens of contrast / edge enhancing and de-blurring techniques I recently noticed a pattern within the obfuscated area. The image consists of 15 x 15 = 225 squares of blocks of 8 x 8 = 64 pixels each. There are 15 horizontal "rows" of 15 squares and 15 vertical "columns" of 15 squares.

The software I used added artificial light to the smudged object from top to bottom at an angle of 3.6 degrees and by going through each row / column and repeating this process 15 consecutive times I was able to literally remove "one layer" of grey pixels of the smudged object and this grey layer was replaced by "one layer"of darker pixels that seem to come forward from the background.

Than I performed the same operation 15 times from left to right but directed the artificial light on the smudged area at an angle of 19.8 degrees.

I repeated this process 33 x 15 = 495 times in both vertical and horizontal direction and ended up with a CLEAR picture without any smudge. I am still working on the analysis and processing of other images with similar obfuscations so I cannot tell you yet what I found.

Are there any "ray tracing" experts who like to share their expertise and answer my few questions?

Is it really a form of ray tracing or did I accidentally used a different technique?

Thanks for your interest,

Greetz,

Sander

edit on 17-2-2012 by 1967sander because: vc

edit on 17-2-2012 by 1967sander because: vcb

edit on 17-2-2012 by 1967sander because: cvcv

Raytracing is a lighting technique used in CG/3d rendering. The ray of light is traced directly to the texture on the 3d model then lights the texture that way.

Basic old school pixel shaders are primarily used right now.

I found this thread a little late..I have no clue about what you are referring to but I do enjoy your threads. And you kinda of have a following by other members in my family now. Although only one other has become a member of ATS.
One said ...many people on here are just too rude and know it all., and he never came back!!
I hope what you are working on turns out to be taking those awful smudges off some of the Moon images. That would be great if that could happen! I guess that is too much to ask for, since it looks as if the smudges destroy the area it covers.
Keep going...you doing more good than you know......

Originally posted by ellieN
That would be great if that could happen! I guess that is too much to ask for, since it looks as if the smudges destroy the area it covers.

Any change to an image that changes the values of the pixels is impossible to reverse, unless we know exactly what happened to each pixel, like "one had it's value reduced to 50%" or "this one had it's value increased by 20".

Ray tracing an object into an existing bitmap or Jpg or any image is very easy.. In the 3D modeling or rendering program, one could set the entire rendering background to a existing image, and then model in an obect or objects and render the new image..

Doing this once could show the operator any needed adjustments needed to make it look natural, and then just re-render it after the fixes..

I know 3dstudio can do this as well as lightwave and others.. Vue software can also do this.


All attributes like blurring, transparency, and materials of object could be completely customized to match anything..

It would very likely be a method used in obfuscation of sensitive images in my opinion..

Originally posted by alienreality
It would very likely be a method used in obfuscation of sensitive images in my opinion..

Why, it's easier to do it in Photoshop.