Acoustic side-channel attacks on printers
Abstract We examine the problem of acoustic emanations of print- ers. We present a novel attack that recovers what a dot- matrix printer processing English text is printing based on a record of the sound it makes, if the microphone is close enough to the printer. In our experiments, the at- tack recovers up to 72 % of printed words, and up to 95 % if we assume contextual knowledge about the text, with a microphone at a distance of 10cm from the printer. After an upfront training phase, the attack is fully auto- mated and uses a combination of machine learning, au- dio processing, and speech recognition techniques, in- cluding spectrum features, Hidden Markov Models and linear classiﬁcation; moreover, it allows for feedback- based incremental learning. We evaluate the effective- ness of countermeasures, and we describe how we suc- cessfully mounted the attack in-ﬁeld (with appropriate privacy protections) in a doctor’s practice to recover the content of medical prescriptions.