Hacker News Reader

Comment by deckar01 | Hacker News Reader

You could make it much harder to detect by synthesizing a unique video with a DNN and hiding the data using traditional stenography techniques.

View on HN

I think that video compression might make this not a viable technique. Artifacts would destroy the hidden data, right?

That is what redundancy and error correcting codes are for. It will reduce your data density, but I am sure you can find parameters that preserve the data.

Couldn't you also embed data through sound? Upload a video of a monkey at the zoo but you insert ultrasound with encoded data.

something like this but far more mundane

https://www.youtube.com/watch?v=yLNpy62jIFk

> but you insert ultrasound with encoded data

Others in these comments have also suggested steganography in both the video and audio streams. The problem with that is that when you retrieve a video from YouTube, you never get the original version back. You only get a lossy re-encoded version, and the very definition of lossy encoding is to toss out details that humans can't (or wouldn't easily) perceive, including ultra-sonic audio.

It might be ridiculous, but how about uploading a computer-generated video of a human saying 0 and 1 very quickly, to encode binary file.

Or better yet, the file could be one third the size if the human says the numbers 0 to 7.

Compression will limit the bandwidth of a given frame but you can work around it.

Some forms of DRM are already essentially this, compression - and even crappy camera recording from a theater - resistant DRM that is essentially stegonagraphy (you can't visually tell its there) exist.

EDIT: "compression resistant watermark" is a good search phrase if anyone is curious

Unless you tuned the NN on the files you get back from YouTube, so that it learns to encode the data in a way that is always recoverable despite the artifacts.

Reply on news.ycombinator.com