Researchers from the University of California (USA) have demonstrated a method by which a gaming computer mouse can be turned into an acoustic monitoring tool. Thanks to the use of ultra-sensitive optical sensors, the device is able to register microvibrations of sound origin transmitted through the surface on which it is located.
Studies have shown that the use of ultra-sensitive optical sensors allows a computer mouse to capture sound vibrations that propagate over a surface and convert them into digital data. Analyzing this data allows you to restore the content of conversations of people in the vicinity. The most vulnerable in this context are mouse models with high-resolution optical sensors – approximately 20,000 DPI and above. Due to their increased sensitivity and high polling frequency, such sensors are able to register even minimal acoustic vibrations of surfaces caused by sound waves, including human speech.
The attack is realized after malware enters the computer. In this case, it is not necessary to use sophisticated virus mechanisms – it is enough for the user to install a program that quietly collects high-frequency data, and the user will not even realize that this program will collect such information from the mouse sensor. For example, just playing a game, or editing a photo or video. The raw signals collected in this way are subsequently exported from the computer for further analysis and restoration of acoustic information.
The received data is processed in several stages: first, digital signal processing methods are applied, including a Wiener filter that reduces noise and extracts the useful signal, after which the information is further cleaned and improved using a neural model capable of recognizing speech patterns. This approach allows the system to extract speech fragments even from significantly noisy signals and partially restore the content of conversations. According to the study, the accuracy of speech recognition within this method is about 42-61%, which indicates the potential use of a computer mouse as a covert acoustic surveillance tool. And if AI is added to this process, the percentage of high-quality recognition can increase.
The purpose of this research is not to create a surveillance tool, but to demonstrate a new potential cyber threat and stimulate the development of effective protection tools. The findings emphasize the need for increased attention from computer peripheral manufacturers and cybersecurity professionals to user privacy issues. Thus, the study not only reveals the possibility of using a computer mouse as a listening device, but also emphasizes the importance of preventive measures to prevent similar threats in the future.
