Behavicker: Eavesdropping Computer-Usage Activities through Acoustic Side Channel

Computers are widely used for business and entertainment purposes throughout our modern lives. Computer kits provide a variety of services such as text processing, programming, shopping, gaming. have greatly enhanced the quality lives; however, we discover an often-overlooked fact that engaging in computer-related activities may be eavesdropped upon by attacker sniffing emitted acoustic signals from keyboard mouse. The activity eavesdropping via side channel has lower requirements terms hardware instrumentation is easier to implement real-world applications than other attacks been presented previous work. In this paper, design system, namely, Behavicker, validate feasibility kind attack. Unlike conventional recognition, Behavicker infers high-level computer-usage with semantics-preserving multiscale learning scheme, based on recognition basic mouse events including left click, right middle scrolling up, down. Real-world experiments show can recognize six interaction accuracy 88.3% infer 82.7% indoor environment.