No Seeing is Also Believing: Electromagnetic-Emission-Based Application Guessing Attacks via Smartphones

Ji, Xiaoyu; Cheng, Yushi; Xu, Wenyuan; Chi, Yuehan; Peng, Hao; Zhu, Zhuangdi; You, Chuang-Wen; Chen, Yichao; Qiu, Lili

doi:10.1109/tmc.2021.3092209

Cited by 4 publications

(5 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Asonov et al [11] observe that the sound of keystrokes differs from key to key and build a supervised learning model to recognize keystrokes. A series of following studies are then proposed to snoop keyboard input via acoustic side channels, by combining different techniques such as time difference of arrival (TDOA) [2][3][4], the language model [13], and a dictionary model [12]. Our work is inspired by previous efforts on acoustic side channels.…”

Section: Computer Eavesdropping Via Sidementioning

confidence: 99%

“…Activities with softwares . Then, we average the first quartile of G 3 3 and the third quartile of G 3 3 as initial threshold and fine tune it to select an optimal one. It is to be noted that, since thresholds here reveal characteristic of distribution patterns of interaction events, they are more independent with users and scalable to different behavior performers.…”

Section: Computer-usage Recognitionmentioning

confidence: 99%

“…Someone may wonder what benefits can an attacker obtain from such a kind of eavesdropping, especially considering that previous works can infer keystrokes [2][3][4]. We admit that keystroke recognition can obtain more fine-grained information compared with our high-level computer-usage activity eavesdropping.…”

Section: Introductionmentioning

confidence: 95%

“…Traditional privacy eavesdropping methods require installing malicious software on target desktops which can be easily detected and blocked. However, researchers have revealed some more imperceptible eavesdropping attacks in recent days which steal private information during using desktops with various sensors such as microphone [2] and magnetic sensor [3]. Most of them focus on recognizing keystrokes and infer the input contents of an attacker [2][3][4].…”

Section: Introductionmentioning

confidence: 99%

“…However, researchers have revealed some more imperceptible eavesdropping attacks in recent days which steal private information during using desktops with various sensors such as microphone [2] and magnetic sensor [3]. Most of them focus on recognizing keystrokes and infer the input contents of an attacker [2][3][4]. Different from them, we investigate another unexplored eavesdropping method which infers computer-usage activities on a desktop from the audio recorded by a smartphone placed nearby.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Behavicker: Eavesdropping Computer-Usage Activities through Acoustic Side Channel

Chen

Lin

Liu

et al. 2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

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

show abstract