Using AI to hack screens through HDMI emissions
Cette publication existe aussi en Français
Researchers in the US have used deep learning to recreate digital video displays from the emissions from the HDMI cables.
The team at Cornell University analysed the electromagnetic waves that unintentionally emanate from the cables and connectors, particularly HDMI.
Compared to the analog case (VGA), the digital case is harder due to a 10-bit encoding in HDMI that results in a much larger bandwidth and non-linear mapping between the observed signal and the pixel’s intensity. As a result, eavesdropping systems designed for the analog case obtain unclear and difficult-to-read images when applied to digital video.
The team recast the issue as an inverse problem and train a deep learning module to map the observed electromagnetic signal back to the displayed image. The proposed system is based on widely available Software Defined Radio called TempestSDR (above) and is fully open-source, seamlessly integrated into the popular GNU Radio framework.
This approach still requires a detailed mathematical analysis of the signal, firstly to determine the frequency at which to tune but also to produce training samples. This saves time and avoids the need to obtain these samples, especially if several configurations are being considered.
The team is working on improving the average Character Error Rate in text, and the current system improves this rate by over 60 percentage points compared to previous available implementations.
They have shared the dataset we generated for training, which comprises both simulated and over 1000 real captures and are also looking at countermeasures to minimize the potential risk of being eavesdropped by systems designed based on similar principles.
www.cornell.edu; https://arxiv.org/abs/2407.09717
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
