18
Conrad
1y

Idea: Hiding a 3D object in an image by making a list how often every color appears and then displaying that as density information in a 256³ cube (aka using a 3D histogram to encode a 3D object)

Comments
  • 5
    I wish I could understand
  • 2
    As far as I understand you use a 3D object (maybe an STL-file from thingiverse), slice it in 256 layers (z axis, will become blue color values), convert each layer to a 256 pixel monochrome bitmap (x & y axes, red and green color values). Then you add these RGB values (256 each) to a known multicolor bitmap.

    Everyone who knows the original multicolor bitmap can subtract it and recreate the shape of the 3D object in a low resolution (256 steps per axis).

    This will work better with HDR images though.

    Did I got this correct?
  • 4
    Just a means of encoding data.
    You can make data look like anything you want if you put in a little effort, even pictures. Especially pictures. Lots of irrelevant data there.
  • 3
    You could easily put that data into 4K image and nobody would even notice there is something in it.
  • 4
    Also steganography is hidding stuff in plain sight. What you said is encoding.
  • 4
  • 1
    Can‘t relate
  • 2
    I’m quite good at concealing my emotions
  • 1
    @lastNick yep thats about it
  • 2
    @robbietoppert it runs in our family
  • 1
    @Conrad wouldn't using the density function lose information, thus making it impossible to decode the object from the map?
  • 0
    yes plz do it
  • 2
    A thought occurs: PNG stego where you just append data to a PNG. It's after the EOF terminator magic, so it should work. You'd just have to encode/encrypt it somehow so it's not awkwardly just... *there.*
  • 2
    Also an idea you could take the data and some image and modify each pixel of the image with data. Change the color of the pixel for example.
  • 0
    @Pickman Of course you're losing information, since you only have a 256³ resolution. That doesn't make it imppssible to decode though. Density resolution is more accurate, since you can create arbitrarily large images with a lot of pixels, and you can assign however many pixels to a voxel as you like. A voxel whose color does not appear in the image would be empty, and the one with the most corresponding pixels would be solid.
  • 0
    @rutee07 I usually use it like in these memes https://devrant.com/rants/2084377/...
  • 2
    @Parzi Your idea will increase the file size while modifying pixel color values of a bitmap file won’t.
  • 0
    @lastNick yeah but forensic steg-detectors and such stop after the PNG-end magic literal.
Add Comment