3 ms·

How can this differentiate between users though? Can the frequency be replicated?

stephengillie 14y ago
Some of that may depend on precision - how precise is the sensor? Can it pull a fingerprint, so it will basically do a fingerprint scan to authorize door unlocking? I'm reminded of #66 on the Evil Overlord list - My security keypad will actually be a fingerprint scanner. Anyone who watches someone press a sequence of buttons or dusts the pad for fingerprints then subsequently tries to enter by repeating that sequence will trigger the alarm system. http://www.eviloverlord.com/lists/overlord.html http://www.eviloverlord.com/lists/overlord.html
- hrayr 14y ago
  #99 Any data file of crucial importance will be padded to 1.45Mb in size. Ah, sign of the times. This list was compiled in 1996.
raphman 14y ago
A persons 'capacitive footprint' depends on skin moisture, clothing, anatomy, etc. This allows Touché to distinguish certain users from each other in a lab setting. As this approach uses only a single electrode (which is why it is so cool), the data it offers is quite limited compared to more complex sensor layouts. In real life, such a sensor would need to work reliably even if I wear different shoes, have dirty hands, etc. Therefore, its false rejection rate needs to be very low. Which leads to a high false acceptance rate, i.e. several other people with similar capacitive footprints may open my door, too.
btn 14y ago
The authors of Touché did a follow-up paper at UIST this year to study this exact question [1]. In brief, you can train a classifier on the impedance profile of specific users, but there are some significant limitations (impedance profiles naturally drift over time, and accuracy doesn't look so good with a lot of users). [1] http://chrisharrison.net/projects/capacitivefingerprinting/CapFingerprinting.pdf http://chrisharrison.net/projects/capacitivefingerprinting/C...

Some of that may depend on precision - how precise is the sensor? Can it pull a fingerprint, so it will basically do a fingerprint scan to authorize door unlocking? I'm reminded of #66 on the Evil Overlord list - My security keypad will actually be a fingerprint scanner. Anyone who watches someone press a sequence of buttons or dusts the pad for fingerprints then subsequently tries to enter by repeating that sequence will trigger the alarm system. http://www.eviloverlord.com/lists/overlord.html http://www.eviloverlord.com/lists/overlord.html

hrayr 14y ago
#99 Any data file of crucial importance will be padded to 1.45Mb in size. Ah, sign of the times. This list was compiled in 1996.

raphman 14y ago

A persons 'capacitive footprint' depends on skin moisture, clothing, anatomy, etc. This allows Touché to distinguish certain users from each other in a lab setting. As this approach uses only a single electrode (which is why it is so cool), the data it offers is quite limited compared to more complex sensor layouts. In real life, such a sensor would need to work reliably even if I wear different shoes, have dirty hands, etc. Therefore, its false rejection rate needs to be very low. Which leads to a high false acceptance rate, i.e. several other people with similar capacitive footprints may open my door, too.

btn 14y ago

The authors of Touché did a follow-up paper at UIST this year to study this exact question [1]. In brief, you can train a classifier on the impedance profile of specific users, but there are some significant limitations (impedance profiles naturally drift over time, and accuracy doesn't look so good with a lot of users). [1] http://chrisharrison.net/projects/capacitivefingerprinting/CapFingerprinting.pdf http://chrisharrison.net/projects/capacitivefingerprinting/C...