Pixelknot is an Android application that allows users to hide short text-based messages in photographs and share them across trusted channels.
What is Steganography?
The practice of embedding secret messages into a piece of media so that no one, apart from the sender and intended recipient, know that the secret message exists. The newly developed algorithm F5 withstands visual and statistical attacks, yet it still offers a large steganographic capacity. F5 implements matrix encoding to improve the efficiency of embedding. Thus it reduces the number of necessary changes. F5 employs permutative straddling to uniformly spread out the changes over the whole steganogram.
The Guardian steganography standard: we are working towards ensuring that the secret message in an image must:
- Have the original image appear, to the trained human eye, unedited.
- Have the bytes of the image appear, to a trained analyst, undistorted so much so as to arouse suspicion.
- Have the complete message be recoverable no matter how it is transmitted.
Have a secret that you want to share? Why not hide it in a picture? With Pixelknot, only your friends with the secret password can unlock your special message. Everyone else just sees a pretty picture. It’s a fun and easy way to share hidden messages without anyone knowing. Take those pixels, twist them in a knot, and see for yourself!
★ DISGUISE YOUR MESSAGES: Pictures are public, the text is hidden inside. Even a trained eye will think the image is unedited. It’s security through obscurity!
★ FOR YOUR EYES ONLY: Put a password on the secret message to make sure that no one can read it except the person it’s meant for.
★ EASY IMAGE CHOOSER: You can use the camera to take photos, or just use photos you’ve already taken.
★ INVISIBLE CHANGES: Even a trained analyst analyst shouldn’t be able to detect any message. The image bytes should seem undistorted.
★ SHARE ACROSS PLATFORMS: Want to share the images over email, file sharing tools (like Dropbox & Sparkleshare), social media (like Google+ & Flickr) or directly through Bluetooth or NFC? Not a problem! The messages are still recoverable on the other side. We’ll have even more tools (like Facebook) working soon, so stay tuned!
★ AD-FREE: We want your love, not your money.
★ MATHEMATICALLY SECURE: We use the newly developed steganography algorithm F5 which implements matrix encoding to improve the efficiency of embedding and employs permutative straddling to uniformly spread out the changes over the whole steganogram.
★ ATTACK RESISTANT: We’ve launched attacks on images with messages hidden in them using a specialized version of stegdetect, an automated tool for detecting steganographic content in images. In most cases, the pictures have been impervious to attack. We will be including detection in an upcoming version of the application so you can easily test it yourself!
★ ARTIST FRIENDLY: The app features the work of Pablo Picasso. His painting “Girl before a Mirror” from March 1932, to be exact. We hope it inspires you to share beautiful imagery and wonderful ideas.
★ WE SPEAK YOUR LANGUAGE: Or at least we try to. Don’t see your language? Join us and help translate the app: https://www.transifex.com/projects/p/pixelknot/
This is still an experimental application and should not be used in real world deployments. We’re skeptical yet optimistic about stenography usage in the field. Most existing steganography tools are security through obscurity. Detection has become easier, but we’re experimenting with new newer methods. Getting users to adopt the tools means insuring that these tools, are verifiably secure, and also usable. We’re exploring options and reviewing best design practices from the field. Overall, this research and not a definitive study. We look forward to further focus, research, and collaboration in the field.
The Guardian steganography standard results:
- Unedited? Many of the apps produced rendering errors, warped in some way. Otherwise, images did not appear edited in any of the apps.
- Undistorted? Initial tests with stegdetect have had success. Trained analysts have not yet been engaged analyze whether distortion has occurred..
- Recoverable? Certain apps re-render the image when sharing, thereby making it unrecoverable. We’ve tried to integrate this into our User Interface to limit problems.
- Source code is available via our Github > git clone https://github.com/guardianproject/PixelKnot.git
- Come discuss on #guardianproject on irc.oftc.net or #guardianproject on freenode.