Dithercrypt is an approach to visual cryptography that works by jointly dithering several images so that the images are revealed when the output is printed on transparency film and overlaid. The dithercrypt program supports four variants:
-
"encrypt2" encrypts a single image into two output grids. Taken individually, the grids are random noise. If overlaid, the image is revealed. Black areas of the secret image will appear entirely black; white areas will be half black and half white; and gray areas will be more than half black and less than half white.
-
"encrypt3" encrypts a three images into three output grids. Taken individually, the grids are random noise. If two grids are overlaid, one of the three input images will be revealed. Black areas of the secret image will appear entirely black; white areas will appear 1/6 white and 5/6 black; and gray areas will be in between. be more than half black and less than half white.
-
"encrypt4" is like "encrypt3" except with six inputs and four outputs. White areas are 1/12 white and 11/12 black.
-
"steg2" encrypts three images and produces two outputs. Each output, by itself, looks like one of the first two inputs. Black areas are 25% white and 75% black; white areas are 50% white and 50% black; and gray areas are in between. When overlaid, the secret image is revealed. Black areas are black; white areas are 25% white; and gray areas are in between. The 25% cutoff is tunable, but the 50% is fixed by the statistics of the process.
The dithercrypt program will work if given grayscale input, but the results will be incorrect, as it does not currently correct for gamma.
To print the results, you'll need to use a program that will give pixelated output instead of trying to interpolate.
This is a placeholder. It's very awkward to use right now.
MIT