So all digital images are made up of tiny pixels that are a single color on your screen. These pixels are represented by binary 1s and 0s in a bit-map image (.bmp) ... more complex file types compress these ones and zeros into much smaller sizes using various algorithms, resulting in .png .mpg etc. Jpegs lose some information when compressed so we can't use them for the next step.
Since each pixel can represent so many colors, a slight alteration to the numbers (say, the last bit or two) is unnoticeable to the naked eye as it simply tweaks the shade of a pixel slightly. A program/tool can systematically alter the last bit or two of every pixels, spreading tiny pieces of a message throughout the picture, hiding/encoding it undetectably in plain sight.
If you know how the manner in which the data was "encoded" (i.e. is it only in the last bit of each pixel? Or more?), you can then take all those altered numbers, build them back into characters(letters, think ASCII for this example) and decrypt(translate) the result with the password from garbage noise to (hopefully) meaningful data.
That's at least the basic overview / starting-point ... it can get pretty complex from there. I haven't played with stuff since undergrad Comp Sci. Really cool to see it used (for good) now. :)
So all digital images are made up of tiny pixels that are a single color on your screen. These pixels are represented by binary 1s and 0s in a bit-map image (.bmp) ... more complex file types compress these ones and zeros into much smaller sizes using various algorithms, resulting in .png .mpg etc. Jpegs lose some information when compressed so we can't use them for the next step.
Since each pixel can represent so many colors, a slight alteration to the numbers (say, the last bit or two) is unnoticeable to the naked eye as it simply tweaks the shade of a pixel slightly. A program/tool can systematically alter the last bit or two of every pixels, spreading tiny pieces of a message throughout the picture, hiding/encoding it undetectably in plain sight.