Guessing Your PIN Right: Unlocking Smartphones with Publicly Available Sensor Data

Berend, David; Jungk, Bernhard; Bhasin, Shivam

doi:10.1109/isvlsi.2018.00073

Cited by 1 publication

(2 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By using the inertial sensors in a smartwatch, an attacker can track the victim's hand motion, and recognize the content that the victim inputs via keyboards [45] and keypads [31], or guess the password of a turntable password lock [30]. By using the data from a cellphone's permissionless sensors such as accelerometer and gyroscope, the attacker can reconstruct the victim's secret PIN for unlocking the cellphone [9].…”

Section: Related Workmentioning

confidence: 99%

“…If the difference is larger than a constant threshold d T 2 = 10 m, we consider that the two edges do not match and continue to try the next start point. If the two edges match, we traverse the following edges of the new path and find if the similar edges exist in the base graph (line [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22].…”

Section: Algorithm 4: Merge the Base Graph And A New Pathmentioning

confidence: 99%

See 1 more Smart Citation

ShoesHacker

Nahrstedt

2019

Proc. ACM Interact. Mob. Wearable Ubiquitous Technol.

View full text Add to dashboard Cite

The past few years have witnessed the rise of smart shoes, the wearable devices that measure foot force or track foot motion. However, people are not aware of the possible privacy leakage from in-shoe force sensors. In this paper, we explore the possibility of locating an indoor victim based on the force signals leaked from smart shoes. We present ShoesHacker, an attack scheme that reconstructs the corridor map of the building that the victim walks in based on force data only. The corridor map enables the attacker to recognize the building, and thus locate the victim on a global map. To handle the lack of training data, we design the stair landing detection algorithm, based on which we extract training data when victims are walking in stairwells. We estimate the trajectory of each walk, and propose the path merging algorithm to merge the trajectories. Moreover, we design a metric to quantify the similarity between corridor maps, which makes building recognition possible. Our experimental results show that, the building recognition accuracy reaches 77.5% in a 40-building dataset, and the victim can be located with an average error lower than 6 m, which reveals the danger of privacy leakage through smart shoes. CCS Concepts: • Information systems → Mobile information processing systems; Location based services; • Humancentered computing → Mobile devices; Ubiquitous and mobile computing systems and tools; • Security and privacy → Domain-specific security and privacy architectures .

show abstract

Section: Related Workmentioning

confidence: 99%