Computing Magazine
Researches have shown that sensors found in our smartphones can sometimes be misused: as an example some researchers used the built-in iPhone 4 accelerometer to build an app that capture keystrokes.
The idea is simple: a smartphone is placed next to the keyboard and the accelerometer is capable of recording the vibrations each time you hit a key. The app is then able to determine the position of each button according to the vibration.
The program is not really able to detect every keystroke, it simply divides the keyboard into several areas: the accelerometer can actually detect roughly if the pressed key is on the top left, top right, bottom left or bottom right of the keyboard. Next, an analysis is performed in overlapping areas with a dictionary to try to reconstruct words.
When the keyboard type is known (QWERTY, AZERTY, etc..) and the language used is identified, the recognition rate of the text is quite high, with a success rate between 70 and 80% . Obviously there are some problems with recognition, especially when a one letter word is used, but the results are interesting.
The type of the used smartphone has also an influence: an iPhone 3GS, which uses a conventional accelerometer is unusable, while an iPhone 4 which couples the accelerometer to a gyroscope is effective. Overall, any modern smartphone used in an environment free from excessive vibration should succeed to do the same work.
Although this might sound innovative and fun, but the goals of the study were to show that the monitoring method are not always obvious to detect, and to raise awareness about using technology in a responsible and ethical way, while appreciating and embracing your privacy and respecting the others'.
The idea is simple: a smartphone is placed next to the keyboard and the accelerometer is capable of recording the vibrations each time you hit a key. The app is then able to determine the position of each button according to the vibration.
The program is not really able to detect every keystroke, it simply divides the keyboard into several areas: the accelerometer can actually detect roughly if the pressed key is on the top left, top right, bottom left or bottom right of the keyboard. Next, an analysis is performed in overlapping areas with a dictionary to try to reconstruct words.
When the keyboard type is known (QWERTY, AZERTY, etc..) and the language used is identified, the recognition rate of the text is quite high, with a success rate between 70 and 80% . Obviously there are some problems with recognition, especially when a one letter word is used, but the results are interesting.
The type of the used smartphone has also an influence: an iPhone 3GS, which uses a conventional accelerometer is unusable, while an iPhone 4 which couples the accelerometer to a gyroscope is effective. Overall, any modern smartphone used in an environment free from excessive vibration should succeed to do the same work.
Although this might sound innovative and fun, but the goals of the study were to show that the monitoring method are not always obvious to detect, and to raise awareness about using technology in a responsible and ethical way, while appreciating and embracing your privacy and respecting the others'.